 But now we are here in 2023 and we are connected on all kinds of channels. We are trying to get better and better. We are streaming everywhere. And yeah, if you want to share anything about the event and you want others to see it, please use the hashtag FuzzAsia. And if you want us to retreat it or to reshare it on any platform, Mastodon or wherever, please also like at FuzzAsia. We are also on these platforms. Cool. Many are already on the telegram chat. Please use this QR code. Join the public telegram chat here to connect with others. I see a lot of people already posting photos. A lot of things going on. So it's a great place to exchange here with the people on site and just connect. Cool. Yeah, so what else is happening? Okay, so we all know we're not yet back to after corona times. And still a lot of people can't travel or travel is super expensive. Also there are changes in the industry. Some people are changing jobs. Actually, we also had a lot of people they said they can come. They have the funding and so on. And then suddenly they have to look for new opportunity. So that means still a lot of people want to join but they cannot join yet on site. So we are streaming on a lot of different platforms. Right now the stream has already started here on YouTube. It's streamed on Twitter. It's streamed also on Asian platforms here, Billy Billy, WeChat, Huodong Sing Live. Also like with the help of our partners like Kayuanshu and Segment Ford in China. So pretty cool to connect and to keep this running. So who's organizing this event? Of course, it's first Asia, it's the community but we also have like other partners and we're very glad to be back here in the Lifelong Learning Institute that is set up by SkillsFuture.sg. So big round of applause to our host and co-organizing partner. We have also like community projects and other companies on board. For the first time here already with a very big booth you saw it on the ground floor, Open Euler and the Open Atom Foundation. Thank you very much. And yeah, a few companies that have been engaging for years for already for some time. Long-term contributor and long-term partner in MySQL, MySQL and Oracle. We have a Rundec by PagerDuty on board. Nice booth with like green printouts on the ground floor and we have only office here and Google. So thanks to our Gold sponsors. And also some new sponsors. We know Prophecy is a big topic. So VPN is a topic. ExpressVPN is providing these services. YDB is here. Grafana Labs with Angel Hack and McCurry. Thank you guys for joining the event. We look it forward to meet you. And we have business partners. You can find them in the exhibition and connect with everyone as well as educational partners if you're looking like for like changing career or some upgrading of your skills. Please check out our partners in the exhibition hall. And of course, last but not least, like a lot of communities, a lot of companies use them. They are here in the exhibition, but also communities are in the exhibition presenting their projects themselves. So really excited to have all of you on board. Big round of applause for everyone. Okay, so if you're online and if you're watching us through a stream here, then also check out the online exhibition. We have a virtual exhibitions area where you can check out all these projects on the left hand side. Just click on exhibition and you will see much more information and connect with the projects and companies directly. So what's happening this year here? The SQR court can scan it like to get direct access to the schedule. We have many different tracks again. And actually this year, I wanna mention like from was like hardware track. It seems like a lot of people are interested in hardware and chips. A lot of developments will risk five because I can see who clicks here. I'm sorry, like it's private, but we still want to know what people are interested in. A lot of people click on hardware. So pretty cool, like big strong hardware track. But followed also by cloud tracks, databases, so professional services. That's what a lot of people are interested in. And that's why we also have like quite a lot of space for these topics here in the event. Security, of course, like it's, we have a dedicated security track, even though you find a lot of security topics also on other tracks. I mean, everything is related. Blockchain and finance, it's a bit less, but like people really looking into the technology, it's interesting. Inner source, how to bring open source processes into companies. We have a robotics track here, big robotics community in Singapore. Standards, of course, web mobile. And yeah, of course, kernel and platform. So it's, everything is related, right? We have some Linux topics in the cloud and DevOps tracks, like regarding containers, but also like focus on actually Linux and other platforms directly in the platform track. So lot going on. And I hope you also know about our social events in the evening already yesterday. We had a workshop, no, a work, no, not a workshop, yeah. So we had people walked around and had a view in the city. Who was there, who participated? Yeah, was it good? Did you like it? Yeah, thumbs up. Okay, happy faces. Excellent. Okay, cool. So a lot is going on, social event on Friday. And we also succeed to make a bit of gear again and we have t-shirts. So actually a lot of people, as I said, can still not participate. So we dedicated 500 t-shirts to send out to online participants. If you're an online participant, please post a message, ask a question and then share the link with us and you will enter the lucky draw to win one of those 500 shirts which we are sending out. Just to say ahead of time, like some postal services don't operate perfectly, but like in the last year, we achieved to deliver 95%. Sometimes it's a challenge in Asia, but 95% were delivered and we want to repeat that. People like post their t-shirts later on, they really feel connected, so we're happy to do that. And yeah, not everything is about online. You guys are here, more and more people coming in. We have of course also shirts here for you. So if you contribute to a session, if the speaker says thumbs up, there was a person really like he helped me sort of back or like he asked a good question or some good input, then you also have the chance to get a t-shirt here on site and directly at the sessions, Weberators will have the vouchers for you. Cool, so don't miss out the social event Friday evening. Talk to the registration if you need more information. In the past we had Indian dance, but you know, Singapore is a diverse country with many like different backgrounds. So this year we invited Malay dance, it's also Ramadan at the moment. So after the day when people can eat and can be together happily like a nice Malay traditional dance fits in very well. So that's what we want. Yeah, karaoke, photo booth. Yeah, just join us for the social event, get the information at the registration. Cool, so that's all from me. And yeah, I'm interested to hear a bit more about the topic that we have here, sustainable world. We had this big topics about Corona, but also we have climate change, we have a lot of topics and it's all the question, how can we make a sustainable world? And I invite Hon Phuk, founder of our first Asia organization here on stage to share a bit more about us and her ideas about this topic. Hon Phuk, please join us. Good morning everyone. So actually what is the idea? It's already on the slide, I don't have so much more to say. Open source, open technology for sustainability. That's all we want to do, during the conference and during the First Asia Summit. But I do prepare some notes that I want to share with you. First of all, once again, welcome to the First Asia Summit. People in the room and also our online participants who've been following us on the stream. I want to say hello also to our broadcasting team from China. I know that they are watching us right now. I hope that you are all right over there. Let's say I remember a year ago when we first started to talk about the First Asia Summit 2023. Within the team, we were so unsure whether we can make it happen again in person. As you know, after COVID, a lot of tech conference no longer organized in person. But then I'm very glad that after several discussion and then a lot of thinking, we once again can have the First Asia Summit in person here in Singapore. Thank you very much for coming here today, for being with us today, especially those who travel from outside of Singapore. I understand how difficult it is to travel these days. And yesterday I met someone, a few people came to the exhibition a day earlier and they told me that it was their first time ever travel outside of their country. And some people came from abroad to Singapore for the first time. So welcome, new connoisseurs. I'm sure that you would like to come back again to the summit similar to many return faces that I see here today. Welcome back, you know, who you are. Thank you very much for being here. Again, for those who are new to the conference, this is an annual event. So we've been running this event for the past 14 years. And the reason why we are doing this stay the same. So we want to gather developers, technologists and individuals who are fascinated about open-source technology and believe in its potential to chance from the world. In the past few years, four or five years, we have seen like so many significant challenges. The pandemic pays like profile impact on all of us on every aspect of life. And I just want to mention again, our main developer in 2021. Who also never got a chance to travel overseas before he passed away during the pandemic. The main developer of event-y system. So we just want to take a moment to think of him. Arif Jamal and beside the pandemic, a lot of other challenges, climate change continues to be a pressing concern that affects all of us in the region. So we can understand it very well. And political polarization, social unrest, have also been on the rise with many countries experiencing wars, protests and civil unrest. So overall, like the last few years have been marked by challenges, but also by when we have challenges, we also have more lives. We see the potential for positive changes. The team of FOS Asia Summit this year is an open technology for a sustainable world. Would I believe more important than ever before? Open technologies is one way we can have to address the issue that the world is facing right now by enabling us to agree and share sustainable solution that benefit everyone. I first started getting involved in open source back in 2007, and I'm happy that I can continue to be engaged and continue to work with open source community. Open source software and hardware can really create a world where innovation is driven by collaboration and not competition. When we work together, we can achieve great things and create a better world for ourselves and also the future generations. Over the next few days, we will have the opportunity to learn from the most talented and innovative people in the open source community. We will hear from experts from the fields of AI, cloud, cybersecurity, open robotics, kernel, operating system, and many more. Mario already mentioned briefly earlier, we also have workshops where you can get hands on experience, and of course, next to the conference group program, we have the exhibition. Yes, I don't know if you noticed already, we have the, in the center, the photo geometry scanner that's been set up just last night, and more than 40 projects by leading companies in the tech industry and also well-established open source projects. Beside that, there's also a few local, very young, developed community building blocks. So this is a group of local Singaporean students who are really into technology and open source. They are attending a series of workshops here tomorrow in the evening as well. We have community coming from the region. We have a first community coming from Korea. Are you here? Yes, so you are over there. So we have community coming from Taiwan. Cosco, are you here in the room? No, probably at the booth. Then we have also community from China. Of course, anyone from China here? Yes, Kai-Yong-Shu, I'm from China. And of course we have people coming from Japan. So just like a few regional communities, if you want to meet them, I invite you to join the exhibition. Finally, what I want to say is, as I said, so the message is we are here doing what we always believe in technology, open-source technology, collaboration, sharing. We believe that this is the way how can we, how we can be a sustainable world. Finally, it's melody mentioned, but I just want to say again, without the support of SQ Future, SG and Lifelong Learning Institute, it's not been possible. And of course our sponsors and partners, thank you so much for your commitment, for your support to the First Asia Summit the past decade. And thank you all once again for being here. I wish you a productive and enjoyable First Asia Summit. See you later. Thank you very much. So our next speaker will be Mr. Marco Antonio Gutierrez. Yeah, Marco, please come over, sit yourself up. While I introduce you, Marco will talk about how open-source empowers the robotics world, the robotics middleware framework in Singapore. And yeah, let me share a bit about Marco. Marco is a software engineer at Open Robotics. Open Robotics is like a community foundation but also Marco works with companies like Intrinsic here in Singapore. And he has a PhD from the University of Extremadura in Spain and contributed to a number of robotics and AI-related projects like Robocomp and the Point Cloud Library or Open Perception. He's also an organization administrator for Google projects, for example, in Google Summer of Code. Okay, Marco, I see you're already set up, very good. So yeah, we are good in time and here we go. Thank you very much for joining. Hello, thank you everyone. First of all, I wanna say it's very nice to see everyone in real life here after a few years. And I wanna give a big thank you to all of you because 2020 showed us that without you this is not possible. So a big round of applause for all of you for being here. Okay, so yeah, my name is Marco. From the Open Robotics team at Intrinsic and I want to show you today a story of the robotics world. I wanted to tell this story for, because there's a lot of people that don't belong to the robotics world so they're not aware of what's been happening there in their industry. And especially I wanna show how open source and free software was a big enabler and I think it's the main reason that we have been evolving so fast and got us to this point in the robotics world. So this is a small gift. It's a big video on YouTube but this is a small chunk of it of contributions to the ROSCORP visualization of the contributions. And I wanna stop here to give a bit of a reflection on open source. I think the humanity grows as we collaborate together. So if we think about anything that we have we could have not done it without the help of others, right? From a t-shirt to like a pen or computers, like anything, right? And I think through time we have found different ways of collaborating together like enterprises, countries, like there's many, many forms of collaboration, right? But they still post boundaries to this collaboration. Like if you have an enterprise you have to collaborate within the enterprise. If you have a country you're not allowed to collaborate or it's very hard to collaborate with all the countries, many, many of these barriers. And I think force is bringing another dimension to this. So when we bring force to humanity, like we're able to break these barriers and boost collaboration further from these boundaries. And this is a story of how that happened in the robotics world and how that helped really boost the development. So there's three main softwares that I wanna talk about. So the biggest one is ROSC, which is the main framework that we use in robotics. The second one is Gazevo, which is focusing simulation for robotics because without simulation robotics would not be possible. And the third one is OpenRMF, which is a multi-robot platform that we're starting to develop and it's actually getting a big traction and it's very relevant here today because Singapore is actually leading the adoption and the development of this software. All right, so a bit about ROSC. So the problem with robotics back in the day was that in every single lab, so robotics had a story of where it was mostly developed on robotics labs, institutes, research and all that. And then suddenly towards the last few years it's been moving towards industry, right? So there's more industry, there's a bit less research. So at the time it was mostly research labs and then there was a lot of reinventing the wheel. Like you have to build a robot, you have to start from the scratch, you have to build this component, even though if your specialty is just perception, so you have to build every single thing from scratch, right? And that would lead to a lot of teams just focusing on one part and then not even developing the rest. So there was very, very few research teams that would have like an entire robot function. They would say, no, but I have this perception part, then I write my paper and that's all, right? But the robot is never working. So it was very, very hard to get like ready, like a full solution ready, right? So there was this, I would say company, part company, part research lab called Vero Garage that was started in Silicon Valley and they started this PR2 program with this robot and then they created a software for it that it was called ROS. And then they send it to a bunch of robotics lab around the world and then they started spreading the software. It was open source and then all these research labs started doing the state of the art research on this and then this got published online and then suddenly you could just use the software and then put it on your robot and it will just run. And then you could do your research on top of that, right? So these reinventing the wheel stop and then we could all share our best contributions and then just improve it on what is our best specialty, right? And then after Vero Garage, the open source robotics foundation was created to protect the software and then took over ROS and now is still the owner of these projects. So what is ROS? So ROS gives two forming areas, provides forming features. The main one is the plumbing which is the one that everyone think when they think about ROS. So basically the plumbing is in robotics we use this concept of components that talk to each other. So basically ROS will be just that framework. You could think about it like as a communication framework. So basically it's that framework that connects your components to each other. They talk to each other and it allows you to insurface what's going on, to debug and all that. There's a bunch of tools that you can use. That includes the insurfection tools that I mentioned but also like visualization and many other tools that are built by the community. And then also you get like capabilities. Like for example, you get the navigation framework. So you can just, you get a robot and then you put the navigation stack and then it's running. Like literally you can get a robot working in one app and like there's nothing else you need to do. Just install NAFTA and get a NAFTA and it's running. So that is very nice. And there's a whole ecosystem of the community where you can just go to this course. You can go to these community events and everyone's there talking the same language ready to help and ready to collaborate. So it's been more than 10 years of ROS and this is the initial ROS that was created back at Wheelogash and all the distributions that we've been having. But because of the change that I mentioned switching to industry, ROS has become a bit more like the requirements have become a bit more different and there's more need for the product to be ready to market, security and all these things. So ROS was rewriting at some point and now it's ROS too. So these are the current releases. There's a rolling release that is the main one that we use for updating the packages all the time and then we just release it and make that release out of that. So the next one will be coming in next May, which is not there. It's, I don't hear any. There's many, many, many robots running ROS in like autonomous cars, NASA robots. It's a picture of the defunct Baxter. When you go to robots.ros.org you will see an incredible amount of robots that run ROS. Startups, there's many, many, many startups running ROS. A lot of them, there's so many that we don't even know because we, ROS is MIT license or some of the startups won't even say it. But these are some of them with some data of how they got acquired, how much money they managed to raise and all that. If you go to this, there's also this link that tries to keep track of the robotics companies, the companies that are using ROS. So how do we manage this? Like it's incredible. It's an incredible amount of people working together and incredible, it's a distributed way. So there's many different parts that work. So in robotics people usually, the problem is so big that people usually specialize on something. So you might be the guy on perception, you might be the guy on navigation, you might be the guy on control, right? So what we do is we have this place called discourse where we all, it's basically a forum, we just post things there and it's the main way of communicating stuff. And then there's working groups. So usually when you wanna create a working group, you post it on discourse, you say, hey, I wanna be working on this and then you just start working on that thing and you start making a meeting and then people start joining and then you start doing your stuff for roles, you create your repos and then you share community. And then there's a more, a bit more structured part on top of that, which is called the technical steering committee, which is the certain requirements that you need to meet. But once you meet them, you can join and these are the current companies that are part of the steering committee and they basically get together and decide what is gonna be the next things to do for each working group, what is the main, the main important things that ROS needs to move forward. There's a ROS2 logo at the bottom, that means there's three slots in the tier steering committee for community members. So there's three community members that are elected and they're also part of the steering committee. All right, so Gazebo, I wanna talk about the simulation a little bit. I'm not gonna stop too much, but basically Gazebo is a collection of libraries that you can use independently because they have their own value, but when you use them together, they become a robotic simulator. And the reason that it really goes well with ROS and it was meant to be go working so well with ROS is because you basically can just use the same software in the simulation and in the robot, right? So you talk to the simulation from ROS and then the same way you talk to the simulation, you can talk to your own robot. You have to change the very minimal changes and then everything should be working. So that saves a lot of time in robotics where hardware problems can be a major issue when you're running tests. These are some of the distributions from ROS, from Gazebo and now I'm gonna go over the robotics video framework happening here in Singapore. So these frameworks started as a problem that was happening here at Changi General Hospital, which is they had one fleet of robot and they wanted to add another one from another vendor. So that brings a lot of problems because the robots don't talk to each other. The robots might have their own needs, like they need a certain line, they cannot share highways, they cannot share elevators and infrastructure, right? So then they basically called open robotics and they said, hey, can you help us build a solution for this and make it open source? So everyone in Singapore can benefit from it and then we can also collaborate and make it freely available. So what did the open robotics build? There's a solution now that allows you to manage these robots all together. You can integrate your fleet of robots with Adam F. And then you can do task planning and allocation. There's a dashboard where you can select which tasks you need to do and then robots will be selected to perform the tasks. It can manage your fleet traffic. So if there's these problems with the robots that they find each other in paths, then you can then conflict this traffic. It integrates with the infrastructure. So we have adapters for doors and lifts and the work cells, which are basically dispensers robots that will, this is an example of dispensing a bottle can of Milo to a robot. There's a bunch of tools that can with it. So you get a traffic editor, which is basically a way to design the floor that allows you to generate a simulation wall and it also allows you to generate an app graph that you can use for the robots navigation. It has the core, which is the one that takes part on this task allocation, the conflicting traffic and all that and it has a dashboard. This picture is not updated, but it is something like that where you can basically select what tasks the robots need to do and when they need to do it. There's many open source assets that had been publicly made publicly available. So you can find them on appgazebosim.org. These are all the companies in Singapore that are part of this collaboration that are somehow also taking part. You can see Changi Airport, GovTech, CGH, many, many, many people working together. So if you're interested, where to start? There's a bunch of ROS links here if you want to get started with ROS. Same for Gazebo, same for OpenRMF. Basically, you get the website, you get some place to ask questions and then you get some place to discuss like usually discourse and a bunch of GitHub repos. So there's so many community events. Rosconn is the main one. So we run Rosconn once a year and the last two years has been online but we're hoping to get a new one. We did the last year in Japan. Sorry, that was in real life and the next one coming in New Orleans. And there's also many other events that happen around the world. Not only Rosconn, there's the local community events. And I wanted to announce that the ROS Meetup Singapore is back and we're doing it as part of the robotics track here at PulseAsia. So big thanks to PulseAsia for helping us run this. And if you guys want to attend, we're giving a very, very basic introduction to ROS for everyone that doesn't know anything about ROS and still wants to get started into the robotics world, and wants to know what is this ROS about. I'll be introduction from Zero and the rest. So you can attend the rest and still understand what's going on. And yeah, thank you very much. I hope there was... Let's try to make some better robots than this. Yes, thank you. Okay, cool. And yeah, so there will be the ROS sessions and I think people can ask questions then and you will go into detail then at the specific sessions. Cool. So the next session is coming up and I would like to ask Norbert and Alexander to come on stage, set up their presentation please. And yeah, I will already start to introduce you guys. So title of your session is the bumpy road of bringing a machine learning model from development to production part one. Search, re-ranking and development of a model. So there will be part two. That's what I get. Yeah, okay, it's working as the clicker working. Yeah, okay, great, nice. So Norbert, we also know each other already for some time, met on different continents. And yeah, you have also been stuck in one place for too long time. But you are back on the road. You've been to the US, you've been to Europe, you've been back here. And you have quite interesting hobbies. I know one time you told me that you had to choose either become a mathematics professor or a mountain guide. So this is the thing. And I ended up as engineer, right? Yeah, you see. So very interesting. And yeah, nice to meet you again here at the event. And you're working now for Merkari who is maybe not based in Japan, doesn't always know about Merkari. We will introduce it in a moment. Okay, great. So it's definitely an e-commerce platform. And yeah, you have brought your colleague with you, Alexander Zagniotov. Yeah, is that our thumbs up? Thank you very much. So more or less right. And Alex, you also arrived yesterday here in Singapore based in Japan as well. How's your Japanese? How is your Japanese? Maybe level. Okay. So then you can, there's still like probably a few years to practice, but like now you're here in Singapore. It's easy. We speak English. So very good. And you create value by using data and technology to drive business decisions. So it seems like you really focus on business. That's great. And you work previously at ThoughtWorks, a German company. And also like has offices here in Singapore. And you worked at Box Inc. And yeah, great to have you here now with McCurry. Welcome, both of you. And we're looking forward to your presentation. Thanks for your introduction, Maria. Yeah, thanks first to the organizers, to organize for Asia in person again. Big thanks to Hong-Fuk and the whole team. I'm very happy after four years to be back. For most of the time, we were not allowed to leave Japan. Well, actually we were allowed to leave Japan, but we were not allowed to enter it again. So that wasn't really an option for us. So I'm quite happy to be here. We are talking, so this is on the AI track. So it will be a bit less about open source. But actually I think there's one reason because open source is so prominent meeting to the whole industry that you don't have to talk about it. Every word between AI, everyone uses AI libraries nowadays, right? That is the big advantage. So what Maria already said, we have two talks here. This one is about the AI part, but don't worry, we don't have a lot of mathematics slides on it, so don't run away. We have a second part on the ML Ops, step-offs part on how to get an eye system running in a rather big environment, you will see. Okay, so what we will do, we will go over some introducing ourselves quick then about what Mercury is, because I guess most of you don't know about it, the state of search in Mercury, and then a bit of technical stuff about how to improve search results and learning to rank and then key takeaways. So let's start with the introduction. So Alexander here, he joined Mercury like two years ago, a bit more than two, he has a huge experience in all kind of famous places. Actually he put me into Mercury because he moved very close to the place I live in Japan, we call it the Inaka Group, because we are living on the other side, not in Tokyo. And so thanks to him I joined Mercury a bit later. This photo is a friend of, as I colleagues sent me, while I was complaining about doing Google Slides because I'm an old style guy, I use latex and produce PDF, and he said, now that this is how I imagine you when you complain about Google Slides. I thought it is quite fitting. Okay, so first about Mercury, so that is actually, it was nice that Tanfuk brought up sustainability because actually Mercury was founded out of one reason. It was the idea to create something which is called in the Mercury speech, a circular economy. It's about reusing and trying to be more sustainable. That was the original idea of the founder 10 years ago, so it was founded 10 years ago. We have now offices in the UK and the US. It's at the core, we have lots of other businesses going on, at the core is a client to client, so people sell their stuff and to other users. Very easy. And of course, if you look at, so you see here, this is more or less how the application looks now. Sorry, it is all in Japanese because while it's on the Japanese market, there is a Mercury where US application is a bit different, but since we are from the Japanese, you will have to deal with that. So the main way to interact with the application is via search, right? You're searching for stuff, so I often buy stuff from my daughter, like ski boots and this kind of stuff. I mean, they have been used once or twice, I don't mind. So this core functionality is provided by Elasticsearch, another open source project. I guess most people who have a beaten idea have heard about Elasticsearch. Elasticsearch provides already excellent way to document retrieval. And yeah, that was the basic use. So a bit of the numbers we are having here, so we have about 150 billion yen, that's about 1 billion US dollar per year, SNET sales, 20 million monthly users, hundreds of millions of active listing, and that you get an idea about. So we had thousands and thousands of search requests per second. So that is the dimension we are speaking about here. And so when Alex joined in April 21, the state of search was just basic. We just throw the search query with a bit of tricks at Elasticsearch and get results back, and that is then displayed to the user. Okay, that works. Actually it works quite nicely, but I mean, there can be improvements. And if you're in this industry and in the area about search and search improvements, then there are a lot of techniques to improve search results with machine learning. So what we are trying to beat up on this regular text-based retriever, Elasticsearch doesn't do anything special, it's just user-read tokens, whatever it is. So single words of it, and it retrieves the best matching document. That quite works quite nicely. But what we wanted to do better than that are get over this, and that is called re-ranking. So looking into what is the state of the art of re-ranking, I will talk about shortly what re-ranking is, so don't worry. And also how we can be improved over time over this. This is not like a process, it's once done and then, okay, we finished. It would be nice, but there are always improvements permanently. So what is re-ranking? In a simple image, like that is a search result, searching for spots, it's like training trousers in Japanese. And what you actually want is that the more relevant stuff, the more those items that have a higher probability to be sold are higher up in the list. Because while at the end, if an item that could be bought by a user is on the six-thirds page, it will probably not find them. So it's better to move them up. This is what re-ranking is doing, basically. So in more abstracting, this is what you get from elastic search, from your basic text search. And what you want to do is to increase the relevance. You want listing to is more relevant to the user who is currently searching. So that's all about, well, what is the reason? Well, it's of course, well, increase money. I mean, we want to sell more, we want that more people use their platform. So that's somehow what we are aiming at. So the basic setup was all done. So we have the Mercury application here. That's at the center. And here's the index with the elastic search that was already here in 21. That was our basic setup. And then there's how can we improve on top of this by just throwing in something that takes the results from elastic search and then just re-orders them in some way using machine learning. So that is what's developed over the, let's say, last two years while it is in. Yeah. And I think here I pass over to... Hello, hello. Yeah, thank you. So I will try to cover, Norbert, so thank you. I will try to cover the ML side of things and just to clarify what Norbert has said previously. The later versions of elastic search are fairly flexible. You can incorporate some ML models and to do fun stuff during indexing time, for example, natural language processing. But it's rather limited when you want to integrate other signals that would add personalization to the search results, for example, user activity or something else. So that's the previous architecture diagram. So this, I will get back to this. This is a pretty common setup, overly, overly, super overly simplified from a very high level. You have the first phase, which is where you're calling results from the index, solar, elastic search, something else. And then you have the other thing which takes the results from the first phase and integrates some other signals that add personalization. For example, recommendation system often work like that when you, if you use Amazon or use Netflix, that's how recommendation algorithms work. They add more signals to the search results when they, or recommendation list when they recommend yourself. For example, what you did yesterday, what you did in the past, your browsing history, what other users in your area, your age, your, with interest doing. Right. So we decided that, yes, we need some sort of a machine learning approach, but what is that machine learning approach? So there is an area in the information retrieval field which called learning to rank, which is basically a set of algorithms which are supervised machine learning algorithms which help you to apply machine learning techniques to add some sort of relevant suspect to the search result, relevance as it pertains to the user that's browsing, researching. So how to choose the algorithms, how to apply it? So there are many algorithms available and luckily for us, there are open source frameworks that already provide implementation of these algorithms so you don't have to write them from scratch. And the way those algorithms work is they approach the ranking problem differently. For example, there are algorithms that consider documents independently of each other, how those documents are relevant to the query, which is called point-wise. There's the parallelize approach where documents compare in pairs and the list-wise approach. So you have the whole result set returned from the first phase retrieval and all those documents together evaluated in terms of their relevancy. So we went with the TensorFlow ranking framework which is the TensorFlow module which sits on top of the TensorFlow core and the reason being is because Mercari is rather maybe TensorFlow oriented so it was kind of more natural for us to choose this framework but to note, this is not the only framework out there and we just ran with it and decided to give it a go. It's backed up by Google so there is some activity on GitHub around that so we decided, let's check it out. First of all, so how to start? So we took an iterative approach. We created a simple model by choosing a set of simple features and we were hoping to iteratively progress and see how our efforts help the search relevance at Mercari. Now, as I mentioned, we used the supervised machine learning approach where we need to label our data. So how do we label the data and what is the label signal? So the most obvious one is click because when you search for something and you want to preview it or you show interest you make a click but there is a problem with that. Clicks are noisy because human behavior is that that human users just click on stuff and it doesn't mean that click means something is relevant and the opposite is also true. It doesn't mean that when there is no click it doesn't mean that the item that was not clicked on also clicks are biased. Normally, human users tend to click on the top results more than they would click at the bottom results. So for example, if you search for monkeys, if you like monkeys and you get back 120 items of various documents that speak about monkeys normally you would see majority of the clicks let's say top 20, 30 results which means if you have relevant results at the bottom they would never get clicked on because most of the clicks are the top which means when you generate a data set for machine learning the labels would be kind of biased which leads to problems like position bias and then selection bias in the data set where you have this loop where you constantly retain your models on the biased label data set. Also depends on your application business domain clicks may not be a good enough signal. For example, it may be a good signal in web search. For example, when you search something and you Google presents your search results and user clicked on a given search result it can be considered relevant result but in a C2C marketplace clicks may not always lead to a purchase. So users click and preview a lot but doesn't mean that what they click on they would purchase. So when we label the data simple click will not be good enough so we need some sort of other signals that would help us to teach the model to teach the machine learning algorithm what is how the model should learn. Right, so as I mentioned clicks are binary labels which means it's either clicked or not or really not not really and this is not good enough. So we adapted approach to create graded relevance labels which means we incorporate other business events when we compute what should be the label. For example, if user clicked on something started made a comment or liked an item started a purchase process and then purchased it means this user event behavior journey with application it can be a good label. So what our approach is not known like other companies approach this more or less the same way. So depending on the business domain you have to adopt your label strategy. Also of course as a general statement in machine learning what data you give to a machine learning algorithm you will get the output accordingly. So to have features of course it's very very essential. So apart from deciding what should be our label we also experimented with number of different features as we were trying to train model. So to touch a little bit about the metrics and what I said earlier about business domain there are a set of very common metrics in information retrieval domain which called NDCG normalize this kind of cumulative gain. It's a metric that calculates the precision of your relevance applying behavior which means the higher the NDCG score it means users see results that are more relevant to them which means they would click higher on a search results which is what we want but a higher NDCG score or any other metric does not mean that the company is actually making more money. So in terms of the numbers you may see higher NDCG score but sales went down. Now back to you and the key takeaways. Thanks a lot for the details I hope you went overhand by the mathematics well it wasn't that much at least from my side. So a few of the key takeaways that if you want to implement something similar right? We are not a big team we could pull this off in short time for a large scale company so using just basic of the shelf open source technologies you can do here a few hints you should take care if you want to look into this. The first is well as already mentioned bad data in bad data better output and you have to need good structure data and proper engineered features that is actually the most difficult part this has nothing to do with software this has with looking at your users what they are doing and what could be relevant. Cleaning your data of course is necessary and invest a lot of time in quality data labeling it sounds crazy as you mentioned before we have 20 million monthly users how many searches per second of course you cannot manually label that that is not possible anymore so you have to think about good indicators for what could be a sign that the user is engaged interested in a certain that is the main thing you have to do if you want to set up a good AI system then keep an eye on the NGCG so NGCG is used everywhere for search results we have seen it again and again if you actually test stuff business KPIs so your actual company or engagement does not really go along always the NGCG so things have to be careful a bit so the general structure is very general system structured in so for the feature engineering here is like can you I mean if it is possible it is good to include into the user interface something that user can give you a feedback that is the optimum stuff that was not a good search page that was not interesting that was interesting even if you get very few of them they will help you to improve your search results it is something that unfortunately is very hard to do because lots of people get annoyed if it is done badly and so one has to be careful cleaning your data I mentioned already so the same happens with extreme outliers so you won't believe what humans are able to do so we have examples of hundreds and hundreds and hundreds of clicks on items in succession by humans we are very surprised what some of our users are doing so that means you have to actually create your data in a good way good data labeling so as I said this is something where you can actually iterate so it is not that you have to come up with a serious solution for all your problems in the first step we never did this we started with binary very trivial labors and then improved incrementally so this is what we call graded leavens or disks like when liked or when commented so we give it a certain graded relevance level this is a way how you can start quickly off off-the-shelf system to provide re-ranking stuff yeah I mentioned already the NDGT blind side so it doesn't mean if the NDGT value is good that you get really a serious improvement just be careful for that yes so a little bit more on the blind side in terms of the number itself the metric might be high but if your first stage retrieval your index gives you absolutely horrible results as it matches the query to the documents which are indexed so in terms of the NDGT it may be high but the results still not relevant to the user because you could have just ranked poor results so your metric goes up but your users are still unhappy that's where the blind side is don't just look at the number and think you solved your element solution is working great okay last but not least there are lots of conferences only related to search in AI there are a lot of things we can do unfortunately we are running out of time I just saw the slides a few ideas one could implement I don't see start with that if you want to implement something similar start with something simple but there are a lot of things what is now let's say the ethics of technology or what most people are using then you can progress to this okay time is over thanks for your interest we are open for I think one or two short questions I guess if there are no questions thanks everyone for that actually the second talk by the two here sitting here we go in more details mlops so we split this because we don't have that much time we have a second talk on Saturday which goes into mlops and implementing and discussing but it's off the shelf stuff I would say nothing specific thanks yes Marco yes because they actually purchase something yes so Marco the question was I will just repeat it for everyone else we mentioned that there is a very weird user behavior and actually user or robots so yes of course this is a huge problem because we all know that 90% of the traffic of the internet is consist of bots we have indication that they make purchases so that could still be a bot but yeah we identified like very consistent clicking behavior pattern from our data it doesn't look like it was a human so and there is like huge amount of clicks so there is certain indications which will help you to identify the bots and you need to clean that out when you prepare your training data set okay I think time is out thanks a lot for your attention and yeah we have to sponsor people if you have questions we around so I have a practical question your second session can people bring their computer and then also like you do some hands on with them is it possible the thing is like we just talked about AI we also have a PhD in AI from Nanzhang Polytechnic here so yeah really great we have your bot and you work a lot with communities to bring them together with companies with projects and we hear more about this now so what should I talk more that's what you talk about so welcome to you thank you very much for joining us so hello everyone I'm Jackie and I'm based in Singapore I understand that there is a lot of flu in over here so a big welcome to all of you to Sunnyside Singapore and I hope that you all had a chance to eat some of our finest food in the hawker centers around but don't worry if you need more recommendations feel free to swing by the booth that we have and I will tell you about all the greatest and yummiest food to eat so I'm from Injo-Hack so I'll just talk a bit about what Injo-Hack is in a bit today I'm going to talk about how we can engage OSS to engage educate and empower developers it's quite mouthful and I feel like I've bitten more than I could chew but it's okay let's see what we can do in the next 15 minutes anyways so before I go this is what my team needs me to say I'm deaf and knife on my throat right now so a little bit about Injo-Hack you may or may not have heard it it's quite an old brand it's been around for the past 11 years or so and it is the world's largest and most diverse global developer ecosystem we have around close to in fact this is kind of like old statistics we have more than like 300K developers all around the world and as its name suggests Injo-Hack we started off focusing on hackathons and to date we have more than built on our events and we have a presence in over 160 countries I hope that some of you might recognize the name supporting us is a pool of global ambassadors all around the world on campuses we can like something that we're trying to start up again but yeah it's very commonly focused and a lot of our events are run quite I guess in a used over decentralized manner but everyone does their own thing within their own community okay a little bit about myself okay that's me right over there I may or may not look like it and a little bit about myself I started off as a scientist in biology and then I just combined a bit of AI research for my PhD and currently I'm the head of product in Injo-Hack where I run where I do research where I did and prototype new products and programs to engage developers in communities myself I'm a passionate educator even though that's my voice is kind of low right I started off with a startup so I'm a zero entrepreneur I'm also an educator I'm a Jackie of all trades I wear many hats in the recent years I ran a data science edu-type company called UpLevel where we help learners level up UpLevel we chose UpLevel because level up was taken so we we help them through project based training and also like apprenticeships so we work with companies we put them in apprenticeships and help them get better at what they do and then I'm going to skip the blockchain part I'm not too sure what the ground sentiment is regarding blockchain for my work in the startup scene I was a world pop 3030 not 30 anymore I've spoken on global stage regarding tech and non-technical stuff I've led projects in machine learning and both blockchain done a lot of simulation work and in my free time I'm an adjunct professor over at NTU who does that right so that's kind of like my hobby teaching so teaching is in my blood I've spent many years teaching now and I think only probably Marco recognizes me amen so I'm not going to talk about what FOSS is exactly right because it would be preaching to the choir I'm going to show you a couple of memes which I think feels like the most representative of the community so I feel like FOSS is a very giving it has a lot of heart it's a lot of thankless work sometimes I feel and they get a lot of flag but this is great also another one of my favorite memes I see they are not really a meme it's I was telling someone that a lot of modern society wouldn't operate without OSS around if you take OSS out of the equation I think society will kind of collapse I think so FOSS is for everyone to play to learn or to do good and I'm going to talk a bit about them quite briefly in each so this is kind of like a breath kind of thing I'm not going to go too in-depth into each part so let's start off with play so what does it mean by play right? play can mean many things here's a question for all of you so what does ICPC ICPC the international college programming contest CATIS and at code.jp have in common so these are what you call online judges they are very popular in colleges where you set it up where you set the challenge up and then the students come in and they work on it a wonderful thing about these platforms is that it's open source if you actually want to run your own contest right if you want to run your own challenge within your community you can always grab the code and then run it within your community so what I'm trying to say here is that let's say if you want to engage your community use such platforms because to me right when you so who here likes challenges who here likes to go for these coding contests alright I don't usually get hands when I ask for it why still do it anyway right engaging community engagement is a tricky thing and you always need some kind of activities to kind of engage with them and having these platforms are wonderful I mean I'm wearing the lenses of a community manager right so if you have all these activities up it's a lot easier to keep them engaged keep them entertained and also keep them up skilled as well play it equals to education and I think one thing that that that's not really statement much is that there's a lot of learning by doing I personally I strongly believe in learning by doing there's only so much you can do by watching videos and reading tutorials on medium or like blogs and whatnot best part is you can modify it freely for your own community right so I've I'm not going to go down that path there's not enough time right so how about learn actually I'm going to spend a bit more time on learn part right because do you know that in a kind of like in a computer science program right you spend around 8,000 years by the time you're done with your four years it's kind of like a rough estimation the numbers may seem a bit different for some of you and for a large part of society right not everyone gets this number of hours of training of course this is based on the usual curriculum in a kind of like university in a developed country right so in countries with fewer resources it's very hard for students to get or developers to get this number of hours of training right so it's actually very easy to get left behind let's say if you're a developer in a less equipped university for example versus say a student from the national university of Singapore the differences in number of training can be quite large of course I guess I can explain a bit what goes into this calculation right 130 credit our module each credit is 10 hours of work per week you multiply it and you get somewhat close to 8,000 hours and this is only for students from how about mid-career switches how about those who come into game a bit late right this is something that I've been trying to solve for the past couple of years in my previous startup and I guess in my educational initiatives right but so that's why I always go back to like using hands-on learning as a way to close the gap between where you are and where you have to be not on par but to have the same kind of experience as regular developer who come out of a CS program so I think I just tried to keep this as general as possible because I wasn't too sure who was in the audience so for example this could be streamed online to aspiring developers out there you are a leader in your community or you are mentoring people there are many things we can do today actually to close the gap and OSS is kind of like the way to go right I guess this list gets updated every year but generally if you the idea is to have this awareness that hey actually there are a lot of these programs out there that you can use so I mean Google is the most famous one right I mean who hasn't heard of some of the people outside the community like this so one thing you get paired up with in open source organizations and then you know you get mentored over the summer right so but if coding isn't your thing there's also the summer of documentation because documentation is also something very important this is also good because on top of training developers we should also be a generation of technical writers those who can write and those who can code concurrently now if this is also something useful like for example good first issue this is what it does is that it kind of collects the data from all the different projects and identifies the good first issues that beginners can work on however if you are in a web free space you have kind of like the equivalent in the web free space let's say if you want to work on blockchain projects instead you can also head to places like learnweb3.xyz there's also an equivalent list of projects that you can contribute to so I'm just going to like skip skip skip skip basically this kind of goes on what I wanted to say if you're a senior you should ought to grab some more mentor it goes both ways right because juniors shouldn't be too shy and reach out to people for mentorship finally do good this is another very fast one we do good why do we do good because there's the implication that's as bad okay so in case you're not sure what I'm showing here's a quick TLDR about what open AI Fiat School is about open AI is supposed to be open but it's not open so it's kind of hard right because AI you see in the news you have a lot of alarming and everything I'm just going to like oversimplify but the idea is that the open source community they are kind of like fighting back right not fighting fighting but they are in response right they are working on their own open source versions so here you have a crowd source data set and then you also have something called shared GPT where they interact so this is an extension that you can install in your Chrome browser where you can just use shared GPT and then upload the conversations you have as training data for okay wait so share GPT right now a quick landscape of how large language models are like in the past three years even though there's been a lot of models released right now these days we are actually focused on the very narrow part of the LLM development so just want to talk briefly about YAMA so YAMA is a model released by Meta or Facebook right and Stanford actually took that and used shared GPT's data to turn YAMA a general purpose LLM into something more equivalent and similar to shared GPT Alpaca so I think we'll see a new generation of animal themed name models it's cute right we used to have things like bird and then you have Ernie and the Sesame Street characters now we'll have animals I love it and it's a constantly developing field and the thing is like where do we go next with regards to doing good with open source I actually don't know what but here's a plug if you want to do good and heck so my colleagues are laughing right now we are organizing something called Hack Singapore it's a pretty large hackathon event I encourage all of you to give it a try I believe it's virtual virtual no it's not virtual it's virtual it's over a long period of time where you can build stuff to do good so one of the themes over here is actually doing good and that's something that we believe in on top of that we have things to help developers level up on discord so two major things that we're running right now one is a monthly code challenge to again hands-on practice to get better this one is data visualization next one is algorithms and the other one that we're doing is what we call a content bounty where we encourage developers to write technical technical content so it could be code guides could be anything so it's kind of like while you learn so in this particular case it's cello right so it's a stable coin it's a stable coin that itself is a separate conversation but developers can then learn the cello ecosystem and then write content about it so anyways call to action if either of it interests you please join us on our discord server and also check out our HEC SU website if you can't catch the QR code in time don't worry we have a booth setup in the hall feel free to swing by to chat I'll be here all day and on this note thank you everyone thank you organizers for having me have a good day ahead okay thank you very much I realize you have a lot of animal photos in your presentations I will say relax thank you I'm Wang Jianmin you can call me Jimmy I'm a software engineer SCS is a leading research institute in China focused on computer sciences but today I will not talk about some complicated research I'm excited to talk about with you I'm very proud to be a part of it in the past three years I will share sorry I will share with you six stories that I experienced in OpenUla community to see how communication plays a crucial role on it that's from the beginning it's December 2019 the OpenUla team come to Beijing to access to talk about cooperation under the vision of OpenUla but it's not a beautiful vision or some good goal to make me decide to join the community it's because the resonance of engineer especially Linux engineer when we start the laptop the first thing we do is not to click mouse or to touch the screen my daughter 6 years old she loves to touch the screen the monitor and we prefer to use keyboard to open a console window with pure black background and enter command right? I see the same action from engineer of OpenUla I see peers same guy I think it's this type of resonance to let me enjoy doing contribution to OpenUla and also let us have better communication in the community OpenUla is just three years old what's the vision of OpenUla OpenUla is what a reliable open source open system that unleashed diversified computing power for a sustainable future as we know we learn from university open system manages software and hardware resources it's about 15 years ago when I start to do some research on OS my teacher Chen Rong told us open system is an accumulation of historical functions in the past 50 years we have a huge and rapid development about software and hardware from first batching it's batching processing multi-tasking into mobile and now to the AI but we think so the open system must follow like the traditional open system focused on supporting the standardized application on particular hardware device we think in all days open system should play a more important role on new hardware technology for cloud computing edge computing and also computing with GPU and NPU to unleash device computing for users so it's a I think it's a big goal, it's a huge goal a lot of things we need to do first thing we need to do is about architecture I think this will be familiar we have many architecture today in the world and in open user we are not only about more architectures it will be very glad that we have engineers from CPU manufacturer like Intel like Kunpeng to come join this community and work together to integrate the newest feature of CPU into open system open user so this is the one thing we need to do a lot of things about it but later we will have more chance to talk about device for computing but let's talk about another one it's a very important one upstream upstream first is the most principle in open user community because upstream project is why open user and other Linux distribution could exist so we pay a lot of attention on relationship with upstream it involves contribution integration and support for example for contribution the kernel CIG in open user many of them are from Huawei kernel team they did the most contribution to Linux 5.10 development and such as Rix 5 CIG they are my colleagues from LCS they also did a lot of contribution for upstream project like Mozilla or LLVM to support Rix 5 so this is such a thing and about integration one years ago all maintainers worked together with linear community and open stack community to finish integration between open stack and we also have ROS 6 we would like to do more integration between them and we also provide support for upstream use of ARM and Rix 5 to upstream project to optimize so this is communication with upstream community the fourth story is about communication in this community two weeks ago I was in Tianjin China we have a TC meeting we have a long talk about what's the next kernel version and what is the next open-oiler long-term support addition it looks like a simple choice but we need to talk with Rix from kernel, compiler release and we also have engineers from Intel to talk about their plan there and sort and we also have engineer from other distribution so this is a suggestion for us we review our timeline review our compatibility policy the coming feature and also users feedback it looks like a simple choice but before we decide before we make the decision the more important is communication because communication makes us let everyone know each other's thoughts we can identify potential issues so I think communication make this community more reliable and sustainable so based on this communication in community we have many engineers driven innovative project around kernel, around robotics, around such as development flexible development and DevOps and this is and then we talk about file one do I have students here? students I think I met some students yesterday in the walking so this is for you the students is our future for open source this story also is from 3 years ago exactly in April 2020 at that time I stayed in Singapore and my colleague Xu Sheng and friend Ma Quan Yi they were in China we talk in an online meeting we talk about how to involve more students in open source project so we think about the idea of the program with some code so we talk with our friends in some open source project and we are very creative so we draft a plan in two days to meet our boss in SAS and OpenYOLA so luckily at that time SAS just run a plan called open source promotion plan so same same goes so luckily the program is approved smoothly students to do contribution for open source project I'm very proud about this one from the first year we set this program global because we shouldn't close the door for communication otherwise we instead we should encourage our students and engineers to talk with international friends and here I will give thanks to some open source community and I'm happy to see the founder of here and thanks the organizer actually this is my first time meeting face to face with Tide and also other founders of the community and 2020 with about 124 community and 100 students from 19 countries included two students from Singapore last week we just announced this year we will support 133 community and also we see some familiar logs in this summit so welcome to the summit and I would like to talk about the last one last story it's about my research institute so why did the ISAS draw in open source community because we believe open source have become the cornerstone for software development we also believe when we participated in open source community contribute to open source and promote open source we will have more good research achievement around open source recently we focused on open source software supply chain and research hope if you are interested about it let's talk about it and so after three years this slide is my favorite because there are so many numbers after three years open oiler we have 800 organization members in this community and we also have 106 special interest group in this community and this number represent the powers of open oiler so and we also pleased to know that in the past three years there are many companies they got based on open oiler distribution and I hope that further commercial achievements will promote the development of open oiler and open source so we got the answer how to build an operating system community to build an aerial solution communication we talk with upstream to help us create innovative project and luckily thanks for Asia in these three days we have many opportunity to talk with you about open oiler this afternoon he is one of technical leaders in our open oiler community she will talk more about diversified computing and we also have senior engineer from open oiler to share with you about innovative project in open oiler so this is all complete information and as TC member, luckily this year I will spend most of time in Singapore we hope we can establish local group and organize more technology events Singapore has a great diverse culture and I believe open oiler also can do a great job on diversifying computing so this is my link in welcome to connect so that's all and may the force of open source be with us thank you thank you so thank you very much and we can now talk with you directly here we have a small break downstairs and you also have a booth where we can sit down right next to the catering area short break 15 minutes please share about yourself and about your activities with code without barriers thank you very much for joining your opportunity good morning I think you had a wonderful morning session I caught some parts of it and here I am to talk to you about code without barriers code without barriers was born during the pandemic we saw two trends collide I'm talking about 2020 early 2020 mid 2020s one was that technology was overtaking the world every organization every person had to be tech enabled my mother suddenly learned whatsapp and zoom and everything and figured out how to talk to me and her grandchildren and things like that every business where it was large enterprise where developers suddenly were remote and had to connect or the small mom and pop shops which had some legacy system which couldn't take the load of online purchases or even the vendor on the street who suddenly had to have a QR code and start doing business through that the technology just overtook everything and in the midst of that there was some technology that was actually changing the way we live and work AI at that time the early 20s and now it's generated AI it's going to completely change how we live work build products and so on and so forth so when we looked at it we saw that a few problems arising out of these AI systems so in 2020 actually or rather 2020 I think the Dutch court forbid a particular AI led software which was looking for social security fraud and the reason it became a big issue was that the court and the algorithm failed failed in the court I mean because they found that what the algorithm was picking up as potential fraud cases were migrant workers single women and things like that and that was one case another one was in in many cities in the US facial recognition has now been banned from being used in order one of the reasons was one big case which called out that most of they were using a predictive model behind to see who's a potential risk and they were using a model in such a case that they were looking at who's a second order of circle third order of circle and pulling up people who had some offenders already in that circle and pulling them and in that more you guess were showing up in the US who are showing up the black male first of all and so and then lot of these models were not explainable it was like black boxes so obviously that was banned another tech giant was using AI in recruitment they use 10 years of data to load their models and then found that women were just being removed from the potential list right of the bat any time there was because they were looking at who are hired basis what was on the resume and so anything related to women, female, diversity all of that right of the bat was getting cut off so they actually pulled it back and stopped using that algorithm so what's fundamental among all of the and there are so many more cases right as simple as black CEO on the search engine you will find truck load of white and now a little brown is added because lot of Indian male CEOs are coming in but in any case it's all male CEOs here and there you will see a few women so when you talk about potential role models to young girls and women who are trying to get into the field they go look at I remember particularly a 8 year old girl who looked up to do software coding she looked up and said oh it's a boys game that's all she's seeing is boys playing on that so all of this combined with the second trend which was 1.8x job losses for women in Asia as against men and this is higher in the west during the pandemic and interestingly worldwide there's 1.1 million gap in women coming back after the pandemic into the job workforce and what does that mean for the GDP that's an entirely different conversation all together but so we saw these two trends collide and we went to the communities open source communities out there spoke to them about how can we make them we first went to women in tech communities of course that is the first stop and they said hey we're trying to do a lot but we're not able to make the needle and so we went and spoke to the general open source communities one example is data and engineering Indonesia there's 7000 developers in there very active regular peer apps spoke to the community leaders how many women are there in your community and he's like what nobody asked me that question ever so I don't know but I never see a woman in the meetups and then the next question but why are they not there he's asking back because they said open community why are they not there and so we go back to the women in tech individuals and have a conversation and they talk about male dominated words that are used a speaker a woman speaker treats questions back differently than a male sometimes the men don't understand why and sometimes the women also don't understand why which is why we thought that there needed to be an intervention and an intervention at a scale that would make people move and so we created code without barriers we went back to the drawing table and said so Microsoft's mission is to empower every developer every person every organization on the planet to achieve more how can you do that if you're excluding 50% of the women out there already so we went back to the drawing board and saw that in our ecosystem we had a powerful ecosystem and there are a bunch of customers and partners who are looking for talent and we're looking for all kinds of talent and female talent as well and then there were communities that we work with so we brought the two together created this platform called code without barriers and what it does is it's open to all communities at joint so we have 31 open source communities today including Farsatia women who code is right here girls in tech so a lot of those communities plus the general open source communities those are the data engineering AI develops all of those communities and we provide programs to skill and certify the women that's fundamental that's a given right but more importantly we went to the customers and partners and said hey you need to do something to actually help the women find the opportunities and create diversity in globalization a typical conversation starts with a data science head and you ask them so you're doing AI so what about responsible AI yes fairness inclusion diversity all part of the responsible AI so how many women do you have on the team or how many diverse thought leaders do you have on the team and that would style them so obviously they wanted to so what they did was about today we have about 52 industry subject matter experts on data AI DevOps you name it all technology areas they are ready to mentor the women who have been skilled and certified so we have a running mentoring circle in the afternoon there's a skilling panel and simple my peer will talk about the mentoring circles a little bit more but these mentoring circles are crucial because women are sometimes held back by women themselves and it sometimes starts from home where the mother treats the daughter and the son differently so we have to get over a lot of that which we do the second thing is allyship so this is not going to happen without all the men in this room participating alongside us right women are only half the equation in diversity and inclusion the other half is men because men are already there being decision makers have a voice women sometimes don't have the voice and your support is actually going to make that difference if you stand up and voice for inclusion so mentoring is a big one and our customers and partners are providing the mentors the second one is hackathon so you can skill as much as you want but unless you are doing you're not building the confidence to go and get at the table and say why you have a thought that's more different in what value you bring to the table so hackathons have been brilliant because one example in Singapore itself is there was a student from finance and marketing who came into the hackathon AI hackathon she skilled AI fundamentals and today she's a python developer and that was the first interview she went to and she had the confidence to build five six other projects in a portfolio and then she applied she got the courage to apply for a python developer post and she got it and she was like wow I'm surprised myself because I'm the first one to get placed from my batch and I got into a tech job so imagine the potential lying there so hackathons just build confidence and bring them out the problem statements come from the customers and partners so they are going to use cases like cars some is talking about ranking of cars and things like that so they bring real industry use cases support the women to hike on it and then get them ready for the jobs and internships so the customers and partners bring jobs internships we have got 200 data apprenticeships from Petrinus it is one of the oil and gas major in Malaysia to Prudential Barclays Johnson and Johnson so it's a cross industries every industry is looking to build the talent and today these industry these customers are getting more aware of responsibly building solutions open source AI all of that and hence requiring diversity in it so my call to you is to come be allies and to the women to come be part of it you are here so you're already ahead of the game but bring more of the women around you participate in all of this and build your story and let there be more speakers and more in AI especially we want everybody like in design development testing user policy making so that's the vision and hopefully we will get there so thank you thank you very much and I think you have to get there so I will do a picture, I bring it up before we start so Sandeep you are senior director in MySQL MySQL here in Singapore yes that's right I'm based out of Singapore that's that picture was probably taken 10 years ago so as you can see the product in front of you is far different from the picture up on the screen my apologies for that No, no, no, I wouldn't say that. You look like very, very young, right? We all stay at home in Corona, sports and everything. So look very well in shape. And so how long have you been with Oracle? I've only been with Oracle a little more than five years. So not too long. And all of those five years have been with my SQL. Okay, very nice. Good. So finally we have the chance to meet. I make it on presentation. Yes, Hanfok, you have a... This is the audio version. This is the new latest version. My friend there helped me put it up, so it's all good. Excellent. Good to have you here. And also like you will share the end about the MySQL track and database track that we have here at the event. So thank you very much for joining us. And a big round of applause for Salim. Thank you, Mario, your two time. Thank you. So I have a ton of... I have quite a few slides and I'm gonna try and run through these as quickly as I can. Starting first with reintroducing all of you to MySQL. You know, the product has been around for some time. It's been there for about 25 years. And for those of you that are not really familiar, there's this company called database engines.com, which monitors about 350 databases across the world, ranks them in order of their popularity. They use different parameters. They look at job boards. They look at technical discussions. And they create a list of people's database popularity. Now MySQL has been the most popular open source database on that list, at least for the last five years that I've been here. In fact, and a couple of years ago, it was also awarded the database of the year, which was an extremely proud moment for us. The reason we are able to kind of maintain this lead is because it's also popular with developers. If you look at surveys done to measure engagement with the developer community, whether they are from Stack Overflow or JetBrains, MySQL remains a very popular database amongst the entire open source, you know, crop of databases out there. And all the work that these developers do, it finally results in the fact that there are a lot of innovative organizations out there that are running across, running on MySQL. MySQL is the database behind Facebook. Facebook, as we know, has three billion users, lots of queries, lots of interactions with the database. It's used in social applications like Twitter. It's used in Pinterest. And then on e-commerce side, it's the main transactional database for booking.com, which is booking almost one and a half million room nights a day. So massive, massive, massive scalability, massive use case. MySQL is also the database behind the e-commerce applications of Netflix and Uber. So I guess what I'm trying to say is that, you know, it's even though very light and very easy to use database, it has been used at scale in applications, in industries across different countries. We see a lot of use cases in finance recently because, you know, with the pandemic, one of the things that happened was everything went digital, including currency. A lot of countries are launching digital currency now, a lot of payment processing applications are coming up. Super apps are coming up. And we see old traditional banks pivoting to these new use cases. And we have a bunch of customers in that area that use MySQL. Old school manufacturing companies are also using MySQL to run their IoT applications. And especially in this part of the world, we've seen a lot of use cases deployed by government. We've got a lot of support from government, to use open source in providing citizen services, you know, to their population. So a bunch of different MySQL, open source applications are running on MySQL. And the common use cases that we've seen are, you know, content management, digital payments, authentication systems. And this list has continued to grow. MySQL is one of the fastest-growing businesses inside Oracle globally. And we're really overwhelmed by the number of users that use our product. We see almost about 100,000 downloads of MySQL on a daily basis from MySQL.com. And what this has really done is it has prompted us to kind of like look back and see how do we provide highly available and secure infrastructure, and database architecture to our users because they're building mission critical applications, right? So in the past, database HA from MySQL was a very manual process. We didn't really offer a lot of tools and, you know, for the database engineer to build high available solutions. And there was a lot of customization that the DB had to do. The DB had to think about, you know, user management, configuration, replication, et cetera. And everything was unique to that particular installation, which also makes it very hard to support because then, you know, there's just a small handful of people who know what's going on inside. Now what we've done is we've come up with something called InnoDB cluster. And InnoDB cluster typically has three, at least three nodes, one primary and two secondaries. And when one of the nodes fails, the other one automatically takes over. And HA is natively built into the InnoDB cluster, so all tasks for high availability are done automatically. You can also have one cluster in one region and then connect it with it to a second cluster through, in a different region through asynchronous replication. And that basically allows you to provide a disaster recovery scenario so that if an entire region goes down, you can do a manual failover and move to the other region. There are going to be, sorry, suddenly I've become very loud. There are going to be a lot of, there are five more sessions around my SQL aspects, such as InnoDB cluster later on during these three days. And I would encourage you to attend those sessions by my more technical colleagues who can dive deeper into these different technologies. The other thing that the whole world is focused around is on security. Cost of data breaches is very high. Data in front of you is from a study done in the US where every data breach is costing a company somewhere around 10 million dollars. And a high number of companies have experienced a data breach. So it's almost like you start to wonder, is it a question of if or is it a question of when I will have a data breach, right? So we have to be prepared around that. And then this is a minor plug into linking back with open source software as well. Red Hat does an annual survey around the state of open source in the enterprise. And when they first started doing this survey about four years ago, the number one reason why companies, CIOs were looking at open source was total cost of ownership. It's going to help me lower my bill. Today that situation has changed where the top two reasons are better security and higher quality software as the two main reasons for that. In fact, 89% of enterprise CIOs surveyed by Red Hat actually believe that open source software is of better quality than proprietary software because people are able to look at the code. The code is auditable. Open source companies have also done a tremendous job of providing bug fixes and patches in a timely manner. And in the last four, five years, the acceptability of open source in the enterprise has really increased. On the same lines, MySQL has another version that we call as the enterprise version that's our paid version. So we love all MySQL users, whether you're using our free version or our paid version. But the paid version does come with a lot of additional features and functionality, especially around security that is built into the product. And you kind of like don't have to look at other tools to try and make a really secure solution. I've listed down a few of the features. And then I have some, you can find out more about this on the web, but basically all the key things like transparent data encryption, MySQL audit, firewall, enterprise masking, a single pane of glass to monitor your entire MySQL estate, all that stuff is available through the MySQL Enterprise Edition. Had a very small nominal cost compared to commercial editions of other database software. One other area that we have been focused around is to maintain our popularity with developers and focused on how to make developers make it easy for developers to work with MySQL. And we've been consistently receiving feedback that MySQL you're doing too many updates. You have a lot of new versions coming up, whereas I've got production environments running on MySQL and I can't keep doing updates and patches on a regular basis. So right now our current version is 8.0.31. The next MySQL release that we have will be what we are gonna call as a long-term support release. And that product will have regular bug fixes and patches, but it will be supported for a minimum number of eight years, three years on extended support and five years on premier support. So if you're running an environment where you want you want complete control, you want complete visibility on and not have to do regular changes and don't have to cope with the update and patch, patch update madness on a regular basis, you can pick the long-term support release as your product. And if you're the kind of customer who wants to look at the latest and the greatest update and be able to have the latest innovation, we're also gonna have innovation releases. We're gonna make it easy to migrate between the LTS release and the innovation release, but you can pick and choose what you want, which world do you want to live in. That I think will hopefully give developers the confidence to continue building their applications on MySQL. And then finally, we also created this marriage between the most popular open source database, which is MySQL and the most popular development environment, which is Visual Studio. And we've put all features of MySQL shell in Visual Studio code, so you can get everything that MySQL shell does to manage and configure your database, but with GUI now that we have MySQL shell for VS code. And then finally, we also added rest service architecture. So you can talk to your MySQL database through the MySQL router. It's supported on OpenAuth 2, provides low level security, great way to serve up JSON documents. Lot of innovation, actually there was a MySQL summit back in Redwood Shores just two weeks ago, so we have had so much innovation and I really encourage all of you to come to the different MySQL sessions tomorrow. You also created a launch and operator for Kubernetes. This product is developed by the same team that builds the InnoDB cluster. And it's currently a level three operator, but it automates all the major tasks of deploying InnoDB cluster in a Kubernetes environment. Our hope and not our hope, our long term goal is to continue to develop the operator and move it to a level four operator which will have provide more insight around alerts, around logs, et cetera. So watch out for this space. And I think one of our other sessions tomorrow is about using the operator in a Kubernetes environment as well. Moving on, all the stuff that I've spoken about is so far is stuff that we have been doing on premise. And if you're following cloud databases, MySQL has had some really great innovation with a product called MySQL Heatwave, which has attracted a lot of press attention over the last few quarters. So MySQL Heatwave Cloud Service is a service that's available on Oracle Cloud Infrastructure. It's a 100% managed service. So whatever typically a DBA does, we do it for you so that the DBA can focus their attention on the application side of the house and all regular OS patching, network management, et cetera is done by the service in an automated manner. The Heatwave Cloud, MySQL Heatwave is what we are calling a single database for OLTP, OLAP and ML. As we said earlier in this talk, right? And we showed you that MySQL is extremely popular with social applications, e-commerce applications. Now imagine you have an e-commerce application connected to MySQL Heatwave, which is, as I mentioned, a single database for OLTP and OLAP. Because of inbuilt machine learning that your customer who's out looking to buy something on your e-commerce application will get real-time recommendations on other products that they can buy. Because again, it's a single database, you can run analytics without ever having to move data out of your transactional environment, putting it into single-purpose analytics database. So there's no ETL required at all. Heatwave works with the standard machine learning tools that AI enthusiasts are familiar with. It also works with visualization tools like Oracle Analytics Cloud, Tableau, et cetera. So one single place to do everything. If you built your transactional workloads on MySQL, this is a really good solution for you and you don't have to think about any ETL activity to work on MySQL transactional business anymore. On a high availability standpoint, on the cloud, deploying a HA cluster is very simple. When you open the OCI console and you're trying to provision, creating your DB systems, you just choose the, hey, I want a HA cluster and the system will kind of create those clusters for you. So it's as simple as that. And if you're wondering, how does this product compare from a pricing standpoint with everything else that our competitors have, we ran some standard TPCH queries on a benchmark and found that it's way faster than AWS Aurora. And because it's faster and is already priced cheaper, customers can actually save a bunch of money. So that's for the Aurora and this is with Snowflake, which is a purpose-built analytics database. We show a lot of price dealership there as well. And then initially, when we launched with Heatwave, we just had one particular shape available. Now we're announcing more shapes available on the basis of customer demand, adding more capability to improve price performance, adding more data handled per node as well, all the while continuing to demonstrate price leadership. Now switching gears and talking a little bit about automation and what are we doing on that front? So MySQL Heatwave has something called MySQL Autopilot, which is machine learning powered automation. So when you're starting to use MySQL Heatwave, it looks at the information that you provided while creating the database system and gives you a recommendation, how many nodes you should have, what should be the shape of the node. It looks at your data and figures out how can it load this data quickly in a parallel way in the memory of the Heatwave nodes. And once the advisor has finished creating this system for you, it continues to kind of like monitor in an automated fashion that entire system. If a node fails, it detects that a node has failed. It will automatically provision another node. It will automatically load data onto that other node. And basically automates most of the regular tasks that you would have to do. And after the system is up and running, it continues to monitor it. It continues to check what have you provisioned for, what is your actual usage looking like? Should you be scaling your system up? Should you be scaling your system down? So all these facilities are already integrated in the autopilot. So analysts have been extremely positive with their praise around MySQL Heatwave. I have some quotes here that from IDC may very well be the single greatest innovation in open source cloud databases in the past 20 years. And I would encourage you folks to go to oracle.com slash MySQL and read some of these Analyst white paper, so to speak for yourself. We have a lot of customer testimonials there as well. We have a lot of early adopters for Heatwave. And actually not just early adopters, now people are putting production workloads on Heatwave. And we have quite a few case studies out there. And I would really encourage you guys to go look at those. At FOSS Asia, we have these five sessions where all the things that I spoke about just in one slide, people are going to spend 50 minutes going deep into them and kind of like having a deeper discussion with you on that. So I'm really hopeful that you guys show up tomorrow in large numbers to all these different five sessions and talk to more knowledgeable people than me as to what really makes all these technologies work. And then finally, if you haven't done already, go create a trial account for MySQL Heatwave. There's $300 of free credits. Look at it, play around with it, ask questions on community forums and get started. Thank you. Thank you very much, Sandy Bia. I'm also very impressed by Heatwave. I really like the people that send a lot of money. Say, speak it different in English, different in German. And I never know which way you use to transcribe your name from Cyrillic to, you know, Latin characters. It should be written with K age, so Mi sheev. Ah, Mi sheev, okay. Not in German, Mi heev, no. Yeah, okay, I see. Yeah, so great to have you, but you are based in Latvia, or where are you based? Me personally, in Uzbekistan. So, but I will talk about that later a little bit. We have lots of offices now as well as here in Singapore. Yes, okay, so I leave it to you. Everything is in your presentation. Let me like take off this thing and then, yeah. Thank you very much. Only office, I think like many years, you're already on board and supporting First Asia and you're connected also to many other people here at the event. We're glad to have you here. Welcome, Alex. Hello. Good afternoon, ladies and gentlemen. It's a pleasure to be here with you personally. First time personally, yeah. And to talk about one of the most important issues facing the businesses in 2020s. I mean, document collaboration. So my name is Alex, I'm with only office and I'm responsible for their professional services. For the next few minutes, I will try to take you through challenges of the document editing and how only office can help your team to overcome them. So let's get started. The pandemic has had a significant impact on the way we work. So now the employees prefer to work from home and remote work is now more relevant. So now all remote workers need to be able to work on their documents from anywhere to collaborate on their documents and to ensure the security of their documents. But all of these challenges can be difficult to overcome without using right software. So but with only office, you can ensure effective teamwork regardless of where your team is located. We have identified four main tasks for us in 2023. So first one is security. It is very important. So with this growing number of the security threats, it is important to make the document collaboration as secure as possible. The second one is the document collaboration itself. So we would like to make the software that help you to organize collaborations so everyone should be able to work with their teammates on the documents. Thirdly, usability. We would like to provide an excellent, I'd like to say brilliant user experience. And finally, flexibility and integrations. And this is I think the key point. We would like to make our software accessible for everyone regardless of their platform or their device. Let me tell you a little bit about on the office. What is on the office? It is an open source project focused on secured and advanced document editing. So it comprises the editors for text documents, slides, sheets, PDF files, and of course digital forms. So on the office has office open XML as a core file format. And it also has a single engine for desktop, mobile and web versions. And you can switch from offline to online and vice versa, that's very important. In 2022, we had, I guess three major releases and more than five intermediate hot fixes. So we do like, we do love making releases. We do love fixing bugs and adding new features. That's true. Now let's have a deeper look at the first point we have identified. So the security. Are you worried about unauthorized access to your important documents? Worry no more. Since version 7.2 on the office has JSON web token enabled by default, what does it mean? It means that everything you need to do is just adding secret part to your host application, to your CMS, to your DMS or anything else. And that's all. When I take a security step further, just replace that default key with your own value again. It has never been easier to protect your spreadsheets with only office. Now you are able to protect your sheets and workbooks. But not only that, you are also able to allow editing only for specific ranges, hide formulas, log sales, or I don't know, there are many options. Version 7.3 takes document editing to another security level. Now you can protect your text document allowing editing only certain actions. So reading, for example, filling forms, there are many, many user writes available here and this sophisticated feature gives you complete control over who can edit or access your document. And now a little bit about the collaboration. Introducing only office forms. What is only office forms? Say goodbye to tedious paperwork routine. Just create model document that contains a lot of different fields like text materials, combo boxes, drop down lists, and many, many others. You are able to communicate on your forms with your teammates. Then you can save them if it's necessary. If you need a hard copy, just export it to PDF, that's all. And we, in version 7.3, we have added a new functionality. Now you are able to create and assign recipient roles for form fillet. What does it mean? It means that it is more comfortable to end user to identify which fields to fill out. We also have added some new fields like date and time with various display options, zip code and credit cards. I'm looking for free and easy to use templates for form templates. We have good news. It is only office free forms library. It has lots of different forms and different languages. And almost everything that you need to automate your paperwork process. If you cannot find the right form, just let us know. We will create it for you. Or if you are interested in placing your form into the library, again, let us know. We will add the information that you are. The author of that form, we will place the link to your, I don't know, to your website, to your WeChat or anything else. We have lots of partners who already have their forms in our library. The next very important news, I think it's a new life view mode. Now you are able to open documents, spreadsheets or presentations in view only mode, but you are able to see what's happening in real time. So all changes are visible. There is no need to wait to click on save on refresh button or anything else. Almost the same for on-office spreadsheets. Here, course of display in your option that allows you to see the selections of your teammates marked with different colors. Again, everything for core editing. We know that version history is very important when you edit your documents. And this is why on-office spreadsheet editors now have that option. So you are able to dig into your drafts to look for an old version and restore it if it's necessary. A little bit about commands. Now you're able to sort all commands in the documents to sort by date, by time, by author or alphabetically. We do have two new modes for displaying changes in the review mode. So the first one is when you click and all changes are displayed in balloons. And the second one is when you hover your mouse over the changes, it's displayed in tooltips. What about usability enhancement? So we have updated our fonts engine. This very important right here. Say hello to our half bus texts fonts library. So it allows us to use new scripts. Now we do support ligatures. Now we are able to combine few symbols into a single one. You know what is it? And now we do support many new languages. Bengali or Singhala are supported now. And we are continuously working on that. On-office is all about customization, all about usability. And this is why we do support dark themes. But not only that, we do support dark contrast customizations and we can enable dark mode or light mode depending on your operating system changes. There are lots of changes in the on-office docs. So for example, you can use local files and URLs when working on mail merge. The hyperlinks can be corrected automatically and the work with shapes is being more convenient. We do support all the spreadsheets and what's very important search and replace engine has been updated. It is now more powerful and more comfortable to use. New features in spreadsheets. Just fewer query tables to preview or hotkeys for page space shows. Just fewer of them. And of course editors haven't left behind so we do support now full animations. We do have new tabs for working with transitions. Again, hyperlinks can be auto corrected and all the spreadsheets can be inserted. And the last version 7.3 allows you to insert smart art objects to use Unicode and Latinx equations and what's very important to use watch window. It's an option that allows you to work on your formulas with corrections, with checking the formulas before applying them in spreadsheets. On the office is designed to work on all file types on all formats. We do have Office Open XML as core file format and we do our best to support all objects and all attributes of that format, of the standard. We have created two new formats. The first one is DocXF where F stands for form. That is a template of the file template of the form. You are able now to work on the form to share it. And then if it is ready, you can share it with anyone saving it to ready to fill out all form format. We do also have new viewers for PDF, XPS and Deja Vu files. Here you can use new page thumbnail panel and new navigation bar. And all PDF files can be converted to DocX and all other file types, except for PDF A. I don't know who does use PDF A but anyway. So all slides can be exported to PNG or JLPG images. PPSX files are also supported for viewing. And now my favorite part, integrations. As I said already, we want to make our software accessible for everyone regardless of their platform or device. And this is, we do have more than 30 integrations. Integrations external with different DMS or CMS platforms, electronic learning platforms and many, many others. But the first integration is the integration with our own solution, only office workspace. What is on the office? Workspace is a collaboration platform that allows you to use different productivity applications with CRM, project management, calendar of mail and many, many other matters. And here we use our DocX editors as default document editing engine. It has been updated when talking about integration. And now new integrations. The first one is Moodle. It is very important. I know that lots of universities around the world are using that platform. We do already have it. And three integrations with frameworks like Strapi, WordPress and Drupal. So if you are using one of these frameworks for your website, you are welcome to give a try to on the office. What about Moodle? It is available in the official Moodle plugin store. So all documents attached to Moodle courses can be edited with on the office now. And you are also able to collaborate on them if it's necessary. Of course, we did not forget to update our existing connectors. Here you can see, I think twice, maybe three times per year, we release new versions of every connector. So C file, library, at first cover, lots of them. And we have welcomed new platforms. For example, Tuleep for open source, Agile management, LEMI-Mach-Pranse for organizing French-speaking community and Flink ISO for quality management software. Now a little bit about plugins. Who doesn't love a good plugin that can enhance your, for example, work on the documents? I'm sure that there is no need to explain what is GC, but now we do have integrations with these three handy platforms. Now we are able to organize voice call or video call right within the document editing software. The next plugin is Draw.io. Unleash your inner artist using that plugin, so draw diagrams, mind maps, or charts. Good news for Markdown fans. We have DrMD plugin. Now your text can be saved, can be exported as Markdown. And of course, but not least, Chart.jpT. So here, there is not about the hype around the Chart.jpT. That's just a small plugin for using text generated with Chart.jpT in your documents. If that's necessary, you're welcome to use it, absolutely for free. And all these plugins can be installed from our official marketplace, whether it's marketplace, you can look for plugins, you can install them, you can remove them directly from the editors without closing them. A little bit about new macros. We have examples for reckon with Google search for importing CSV and TXT data, and again, for example, for using Chart.jpT. As I said already twice, we want to make our software accessible for everyone regardless of their platform. And this is why we do have lots of distribution forms and packages. And a good news, only OfficeDocs is now available as a service, as a cloud service. There is no need to host, to customize, to install anything. Just go to on the Office webpage and register on the OfficeDocs. Good news for ARM fans. On the OfficeDocs and DocumentBuilder are available when using devices on ARM architecture. And we are proud that our desktop editors are included into many operating systems, many operating systems. Here are just a few of them. And I'd like to pay your attention to the fact that we have Compatibility Certificate with China's operating system, Kailin OS. Our software can be provided as a service by different service providers like Alibaba, OVH Cloud, or OVH Marketplace. We are glad to be involved into some technological launches. For example, Miniform and Manjaro Linux have created their own mini PC, URM 350, with only Office Desktop pre-installed. And the same for Murena cell phones. They do provide our software in their ecosystem for editing their documents. If you are using Angular or React or Vue, you are welcome to use one of our examples. There are a few available on our webpage. And one good news for developers. We have created a new API class to access documents from outside, from external interfaces. Now you are able to communicate with the content of your document without relying on document editor buttons, just using API methods. A little bit about document builder. We have updated .NET, Dr. Render Library, to make more comfortable working with a document builder if you have application written in .NET. And a little bit about our team and our locations. We now have new offices here in Singapore, in Armenia, in Uzbekistan, and in Serbia. Our team is growing, and that helps us to provide high-quality technical support, high-quality professional services. That is very important for us. Again, we are here in Singapore. We are glad that we have won Cloud Content Management Award in Cloud Computing Insignia Awards. It is an honor to be recognized as one of the most important Cloud Content Management platforms. A little bit about our future plans. We do have lots of plans. We do have plans to add new features. We do have plans to add new options. But we also do have plans to make new products. And this is why we are working now on a new product called On-the-Office Dock Space. What is it? We know almost everything about editing documents, about co-editing documents. And this is why we have created a new product. It is a new way to collaborate on your documents in the rooms called... In the space is called rooms. So you can collaborate with your teammates. You can share your documents. You can invite new users. You can customize the rooms. And what is very important, this product can be integrated into your solution. We do have plans to include end-to-end encryption. And our forms also can be integrated into that product. That is absolutely a new way how to co-author your documents. How to work with your teammates. We do have lots of plans. We do have plans to add right-to-left support. We do have plans to add new languages in the interfaces. To edit documents on mobile devices in a new way, in a brand new way. And to add digital signatures. Again, talking about the previous product on the office dog space, digital signatures are also supported. Thank you a lot for your time. So we have a booth downstairs. If you are interested, just come. We can show you everything you need or you are interested in. Thank you. Thank you, Alex. And how long will you stay in Singapore? Three more days. Okay, great. So yeah, time to catch up and learn more about this. About only office. Been using it for years. Thank you very much. Cool. And before the last one, we have the last session. Okay, cool. So yeah, how long will you stay? Alex, how much time do you have in Singapore? I'm sorry. Sorry. Until Saturday. Okay, that's great. And I heard that you are working with quite a few people in Singapore together. So there will be a hardware track. We had a lot of interest. People wanted to join that hardware track that will be on Friday. And also a few more sessions on Saturday. You will have a longer session there. But we have a short peak in your work right now. So thank you very much for sharing about this. Johan, thank you. So yeah, thank you for having me. And I'm happy to be here to give a short intro to the development ecosystem. And so my name is Johan of Rosine. On Line I Go by Propie. I joined Google in 2011, so a little bit more, 12 years ago. I've been doing a developer relationship there. So I'm a developer relationship engineer, meaning that I'm focusing on improving the developer experience for a variety of Google developer products. I started working on some of the cloud developer products of Google first. Then I worked a little bit on Android too. And more recently I've worked on IoT and hardware things. This year, last year I joined a new team at Google called the hardware Toolchain Team. And my presentation is about what this team is doing and how we are trying to build an open source silicon ecosystem. So my core team mission is that it's to make custom silicon easier to build for everyone at scale, just like software. It's something like, I know many of you are a developer and you're familiar with some optimization file that you can pass to your compiler. And sometimes you go, you max those optimization possibilities and there is nothing that you can do in software to optimize further. So our vision is that we'd like you to go to the next step, to optimize your stuff in hardware. And usually the gap between jumping from a software product is pretty big. And we'd like to make it that as easy as just changing your compiler flag. So we're not there yet, but that's like the vision. So imagine that you have an optimization flag and you can say that you want to optimize stuff in silicon and then give you the hardware design to optimize. That's a translation unit. And in order to get there, we found out that there are missing pieces. So in ICCAD in 2020, the team founder actually released a paper there called The Missing Pieces of Open Design Enablements. He identified like four things that are missing for creating like an open ecosystem in hardware that could strive as much as a software one. So first like open source PDK. So in software world we take it for granted that we have like some open source SDK and library that we can build or work on. And often like work is consistent remixing like a few libraries that we want to produce something original. Like it doesn't really work like this in hardware. Like the very lower level of the stack, which is a PDK, you can think of it as an SDK for hardware. It's a thing that define like the specification of the foundry, of the process of the foundry, the thing that will allow you to manufacture something with that foundry. It's an interface of the foundry. And those things currently require, most of them require an NDA. And so like you can't even start developing our other project until you sign a contract with the foundry. The other thing is like and so we realize that if we want to allow people to do custom hardware we have to have open source PDK. The other thing is a silicon touch and so we take it for granted in software that most of the compiler or debugger or optimization tool are open source. And you can run it on any computer and you can run it in the cloud. It doesn't really work like this. It wasn't really working like this in hardware like a few years ago. Like most of the tools were proprietary. You had like pro-itv license cost to get started. And even more important you couldn't run like those tools really freely on the cloud or on weird architecture. The other thing is like because there is this bottom of the stack with the tooling and the PDK that were mostly proprietary. It meant that basically all the building blocks that people were building on top of those two were also tied to proprietary solution. So there was like very little reuse of like other people and sharing between people design. And the last thing is like it cost a lot to manufacture silicon. It cost like tens of thousands and hundreds of thousands of dollars just to like a simple batch on a very, very old technology. It could cost like as millions of dollars to do something on a more modern process. Meaning that the cost of error is huge. Like if you make a design mistake like it's going to cost you like thousands or millions of dollars. And so it's really pro-itv for people to get into that field because like in order to learn something you have to make errors. And if they say or could you cost you a huge amount of money you're not really incentivized to learn this. And so we identify that if we basically provide open source PDK and source silicon to chain. If we provide an easy way for people to generate blocks that they can share and reuse. And if we provide them with cheaper and faster manufacturing option we can create like a bootstrap an open source ecosystem that could be comparable with software. And so that's what we try to do with that team. So we partner with a few foundry across like the world. So the first one is skywater and we try to convince them to open source a PDK for one of their older 130 nanometer technology. So there is like two variants of the PDK so there is a Sky 1 and more in 30A that's like a regular CMOS process and there is the Sky 1 30B which is as the same process with a rerun cell so resistive run capability added. And that was like two years ago last year we managed to convince another foundry global foundry to open source their PDK for their 180 nanometer technology. So it was very important for the ecosystem because like we went from one to two so the gap between 0 to 1 was big and a lot of the tooling actually started like supporting this skywater PDK and the gap between 1 to 2 was also very big because like all the tooling had to kind of remove the things that were specific to skywater to start like supporting another PDK. And like this year there was also something that was very good for the ecosystem that happened. So we had this foundry in Austria called IHP that released their open source PDK their PDK and open source PDK for their 190 nanometer technology without talking to our team at all. So before like Google kind of went to this foundry tried to lobby them or convince them to open source their PDK the first example of an organic release of a foundry that didn't consult at all with us and that started releasing like their technology. And later this year we hope that we'll be able to release like a 90 nanometer PDK for the Sky90FG process. And so also of the PDK you can run like a torsion on top of it and so we support like many of those open source tools from that ecosystem. There is tools that we develop ourselves. So like for example XLS is a 90 level synthesis tool that Google develop in us. And that we make sure that this tool support like the various PDK that got open sourced. We rely for synthesis on a tool called YOSIS that's very well known also inside the FPGA ecosystem. For place and route we rely on open route and with a flow called OpenNain for the last step which is producing a file that you send to the foundry for manufacturing we rely on magic and Kaliot. And so all these kind of tools can give you an idea on how you can go to a iLevel description that's really close to source code that describe the functionality that you want to accelerate. So in your case like it's a X file that it's a syntax like close to rust that allow you to describe like a algorithm that you want to create hardware for. And then you go to Verilog which is like something that gets really tied to, it's basically it can get synthesized into a netlist that match like the process that you want to use. And then like this netlist get plastered with like the cell on natural silicon dye and that's like what the left and the deaf are about. And then you stream that into a GDS format which is describing the actual polygon that are being sent to the foundry for fabrication. And like all these tools can allow you to like produce this file like using only open source tool. And once you have the PDK and you have the tool you still need a way to manufacture the chip and like I mentioned before it could cost like a 10,000 or maybe like 100,000 of dollars to manufacture chip on those old even of those older technology. So we started for the two years we've been running like an open NPW Shuttle program that basically come from no cost to the community. Like the community only need to produce like an open source design that's reproducible with the open source tool to unroll into that program. And we've been selecting like 40 projects on each Shuttle round to get manufacture. And so like at the end like you get like a board like this with a little custom silicon chip on it and you can like start like validating that the silicon result that you have actually matched the intent of your design. Like every single chip come with a user area that you see here where you can put your own logic and a little like this white chip on the bottom that can help you like troubleshoot your design. And that's like one of the picture of the wafer that got manufacture on one of the first one. So we've seen an increasing engagement from the community on the Shuttle so we run like 9 Shuttle on this program. And like here you can see like we started with only like 40 projects at then 55 and on the last Shuttle we had as most as 150 projects that got submitted. And so it means that we have committed like a pool of more than 700 design that are fully open source and reproducible with the open source tool. So the tool have changed a lot so it could be changing to reproduce like one of the earlier one with like the newest version of the tool but still the source is available like anybody could look at the project and anybody could try to reuse like the bits that are there. And we've seen engagement from all over the world on this program. So that's like kind of a picture of like the breakdown of project by country for one of the latest run that we run like GF NPW0. And so it's not only the United States we've seen a very good engagement from country in APAC like Japan, India and currently we are busy with the community trying to bring up like the silicone for the second shuttle the NPW2 one. And here you can see the various example of people like playing with the ball like that silicone is actually working and like kind of posting like the result online. I wanted to showcase like three projects that we've seen on the second shuttle where we've got like good result. So that one is like from an actual professional like a soft designer from Intel that has like experience in hardware design. And so he did like this little RIS5 microcontroller that's compatible with Arduino and like it has like an SD RAM memory controller on board. It also has a quite SPI like interface and he was able to get like his cheap booting and validate that the silicone was working. Something that he said is that he like he has been in this industry for a long time he wouldn't think that it would be possible to kind of build like a custom silicone directly from his own laptop and being able to manufacture it for free. So someone else that I thought was was bringing up was like GetCat with like the project that they did with their university, from the L.D. Burke University. So what they did is that they did a custom FPGA. So they implemented like an FPGA fabric using the sky weather program using the case of sky weather process. So there is a tiny amount of loot there. It's only 100 loots. But the achievement is impressive because here they were able to run like a custom FPGA logic using an open source FPGA toolchain targeting a chip that didn't exist before now running on the FPGA a custom FPGA fabric. So the amount of things that could fell there inside that stack is like impressive. Like the toolchain couldn't possibly not work like the design could be flowed, like the silicone could also be problematic and like they were able to validate that silicone and validate and here you see like in the middle they have a little FPGA anymore that run on their fabric and that show it working. And so yeah that was like impressive that they managed to get that working on a brand new open source PDK. And the last one is like kind of dial well into the narrative of democratizing IC fabrication. So it comes from an online course called Zero to Azix where they actually instead of using like the whole cheap area for just one design what they did is that they run that course online without really asking us for any permission. And they just entered like a shuttle and not submitted one design but submitted 16 design from the this course into one project slot. And like most of the people on that I think like 12 other people of the 16 were actually first time designer. They never did a hardware design before and they were going through that course curriculum and at the end like were able to submit their own design. And so it also show like how the community can if you give like the community like constraint like for example you can fit like inside that amount of silicone area they will build like their own process their own learning material. And they will kind of adapt to this constraint and try to build something else. And I mean I think it's very impressive what they managed to do here because they managed to get a lot more people on top of this infrastructure that we set up with better learning experience that we could have built. Okay. So another thing I want to tell you is to like please if you're interested in hardware like feel free to join this community. We have a Slack channel with more than 100 people and we have like this website called developer. And I never talk about that where I will actually go inside bringing this tool with you on Friday. So if you're interested feel free to come there. Sorry I went over. Okay. Thank you very much and yeah more on Friday definitely. Also know you work with they will all be there. Cool. We now have a lunch break. The lunch is sponsored by the Many Chinese universities also include the open-source course on the university. They make for and they make community, they make for the conductors. That the totally open-source ecosystem study on the university. And full-time engineers also join the group. How you are managed for open-source development? They support for the open-source community and culture. Chinese big company current big donation for the New Yorkers Foundation and the Dutch Foundation is an engineer who contributed a lot for the open-source. Many companies oppose for them. I think they are very good running of the carbon and hydro-source and western huge companies who are real of the open-source. Current Chinese huge companies also support us. They are also inclusive. If you are an advocate for the Chinese open-source users, why do you use open-source software? Previously, just free, no need to pay is the most important. But now transparency is the most important. Second one is a future. Third one is a saver time that I want to develop. People are understanding what is open-source, what is made for the open-source. What open-source could change the technology. The rest are summarized from the company side to the open-source. Big tech company is the open-source division and a hired professional open-source committer. This is our kind of job for a hired open-source enthusiast by Arirama, OPPO, Huawei, ZZBotcom, and HikeVision. Making hardware companies, software companies, hired professional open-source committer. This is about an example. I'm not a crowd-op open-source head-to-goer now. There are a huge open-source programming office, OSPO, and they need a more higher engineer and open-source advocate. Make good calls and how to make good calls by the community. It's most important for Chinese companies. This is an example of an open-source software by Chinese. It is the most famous open-source software from China in Japan. I saw 40 books about open-source in Japan. And after the line, this is the best of the U.I. tool kit. There are 400 stars on BigHeart. The stars are also very famous in China. This company called BigHeart. After the post-Taipei, they got 2,200 million dollars in 3D open-source. They are home from China. But their branch is Singapore already, New York already, Japan already. And I have seen the presentation of the China open-source conference. He said, beginning of the era of Taipei, calling the time on BigHeart is only Chinese break time. But now, almost 20 hours, 365 days, continuous treatment from all over the world. That's the reason why Taipei is a global software and Taipei is more a mandatory software for the society. That's the reason why they did the company. BigHeart company is providing software as a service platform for the users and still supporting the community. BigHeart model is Huroku, Docker, and Electric Stars. China completely copied that BigHeart model in there. And this is a chart from the Chinese venture capital. Fundraising OSS company and some companies called Unicode already. And Chinese fund focused on open-source. Because we, Kai Wencha, and other Kai's members, are working for the business. If the company is software from China, how to make the market do well? Very, very, very hard. Extremely hard. But the open source, open standard from China are easy to bring to the ground. Many Chinese companies, not only from IP, bringing the more developed standard, bringing the more developed building together. That's the reason why China government and Chinese company focus on more and more open source. This is most important slide we might share. From the Chinese venture capital, they told of current open source 3.0 era for the business. The beginning of open source, open source 0.0, is a typical softwareizer, ggna and relax. All of us know BigHeart model, but just for fun. I really like just for fun. And also just to change the society. This is the most primary important on the path. And second one, gna, relax, relax, high SQL, PHP era. Building the software, building the solution, building the service by only open source software and doing the same thing as the software development. This era is almost the same. As a normal company, use for free, support for the pay, is a version of 0.0. And version 2.0, software as a service era, clandera, component of Kafka, these are mixed community beta, mixed software as a service beta, and mixed for good service by the community. This is a very good way from Apache software foundation, they saw the community over thought. Mixed good community, they contribute and renew the software a lot. That's the reason why we pay for free for the software as a service. Then, now, newest one, open source software 3.0, like MongoDB, Databricks, also BigHeart and iDB, which is not just a cloud native, always an open source software, just a platform as a service model. Scarable, disk-reviewed, cloud service, almost all solution is open source. Very hard to find a proprietary, but we still pay for the software. Currently, community is most important. How developing the software is most important. Primary or open source is most important. If you're making a global community, if you're making a big developing power, if you can hire a million engineers, you can do the proprietary, that's very hard. Good is also sometimes doing the open source, because good people are also very clever. Good is brilliant engineers there, but they know although good is much, much better. That's the reason why they choose the open source. And also the community style, our community is still growing. Also in the foundation is the global community. Trash can remember is far away, Tencent cloud, Antedrupe, Anipada. We are working together with Sony, Erikson, Facebook, Winniar, Red Heart, Microsoft. In the open source style, western people and eastern people working together to make a platform. Currently, open infrastructure is making very good cloud software, expect awesome. Many projects from China, Incubation by Mirage Foundation, Apostos Air Foundation, CloudNate Foundation, Mirage Foundation makes 50 projects already. Just 15 projects is all coming from China. Foundation was there. This is a simple, China government supported the open source foundation. They are ocean foundation. They can make a regulation. They can make a goal for the open source in China. They build the license, integrate open source software from China, and push open source infrastructure from the government, and build the university and the company, and make a role support in open source. This is one of the examples from the Open Atom Foundation. They make the new license, Moorang Software, Famicom Software right as part of too. Moorang PSL. This is almost the same as the past license. But Moorang is China friendly. Because in China, some systems are different from the United States. This license is a good for the Chinese society. And also, the English and Chinese are compatible. This license's purpose means to make contracts from Chinese company to overseas company using the same license for the open source. This license is a open source infrastructure already in the open source infrastructure. And we Chinese making new one for the GDL style license for China. And also open source hardware style. There are a lot of Chinese style copies here from Shanzai. Shanzai company makes a lot of copies. But this Shanzai style is transforming now to the Gronka. Later, we are coming to each other his way. These two are ideas from my part. New Shanzai engineers and entrepreneurs are helping the Chinese technology such as Xiaomi, OPPO, and the Haki layer robot. I try to explain to them where the Shanzai comes from and where the Shanzai has to go out. Shanzai means Chinese style open source hardware. But design change is based on license. That is why open source. We share the design. We share the hardware and we share the IP. This is a function of serial open source but not based on license. Current Chinese company is a transfer to share IP, share product, share knowledge but based on open source license. Currently, more than 550% of risk-fired premium vendors are from China. Currently, China is the most popular risk-fired company in the world. You can see some clients company was there. Rick Kestova is the best of Alibaba. They sold a more 2.5 billion of chips by risk-fired vendors. If you have a true buyers here and there, the price under there keep from China, Shanzai, Jelly or Plutons. These two companies are very huge advocate for the risk-fired. Currently, risk-fired is a cover already. If you bought the Bluetooth chips under 10,000 dollars, probably risk-fired chip from China. Alibaba is the chip design company he had 6 out of 2.5 billion of risk-fired chips. They are open source. No any Shanzai was there. And the rest of 10,000 Bluetooth chips and Bluetooth and Bluetooth speakers are sometimes tired about the Bluetooth IP, but chips and IPs are purely risk-fired. Our company sweet science is an open source hardware marketer. We sold a lot of Raspberry Pi, Arduino, and M5 Star, and also PSLab. Many open source hardware with 5 chips, StarOS and Sensor from China. I spent the biggest from China from 2014. 2014 hardware is ok, comfortable even cheaper than other. But last time hardware from Star from China I never saw yet new type of technology from China, like AI, A6 model and also one chip SOC installed a lot of functions Shanzai transforming to the public now, there again. In this IoT development of hardware for M5 Star ESP32 desktop in screen and battery there are some markets that are currently bigger than Arduino in Japan. Thanks for the attention. Currently this activity in post Asia is very good for that and memorable for the Kaiwen Sha. Three Chinese open source community want more and more to contribute to the world of open source community. We start from Rami, we use a lot of open source and currently we want to contribute to the world of open source community. We think more and more to make and also more and more in Singapore as a society. Marseille, Singapore, Marseille open source society. Thank you, Wakami. Thank you very much. Can we talk? I'm going to ask questions. In Japan Japanese Japanese Excuse me. Can you hear me? Yeah, you talk about something like an Arduino so what is the difference between in which you talk about the difference between the books and the change of the world in Singapore? You have a question about the different society open source in western Thailand? Of course almost severe I know in my knowledge almost all open source license is based on United States even if we find a contract we have to bring that into the U.S. court like in Singapore I keep an eye on the license but you are out of the license they have to submit that contract even if it's a shorted contract we have to go to the U.S. court but in China is the border in China Chinese government is open source Chinese government is open source open source for the Chinese I was just wondering you mentioned a lot of development how is the Chinese government approaching other governments around the world I think that China is a more European service by the government in the Chinese society open source the Chinese government has to proceed most of the big things in the world secret from China very hard to sell most companies in the Chinese market are only in demand Japanese companies even my solution is appropriate very hard to sell from China okay thank you that company pick up American company square Japanese company also using the okay American company square Japanese company also using the Chinese open source in exchange the question is more about what you think a Japan open source communicator from what pirates are doing in China I heard from Japan a lot of techies are doing so and hard to understand the Japanese open source Japanese open source society are the students doing the contribution that most of the world problem in Japanese society the business guy and government is very very far from the deep end you can't understand the open source most important to learn the open source in here China is a better thing because many business guys many government guys more open source and good open source and well thinking and design very very far this is the largest problem in Japan I also want to show you that program the open source is all over the world thank you very much thank you Takasu and we are looking forward to the other section here the event thank you so much for meeting me with us thank you thank you thank you thank you please enjoy together for the possibility after three days