 All right. Fantastic. Hi, everyone. Lovely to have you all here with us for the kickoff of day two of the e-research data skills summit. My name is Dr Tyndale Sumner. I'm a research fellow and consultant at the University of Melbourne. I've worked with many people in this call in a number of capacities in the e-research and data skills space across the country. And I'm a humanities researcher with a particular interest in surveillance studies and literary studies. We've got a really fantastic lineup of speakers today, very cross or transdisciplinary, with some really fantastic and interesting talks around accelerating skills development. We're going to be looking at AI computing services and a really fascinating new online space for cross disciplinary research methodologies. I'll just get you to jump to the next slide, please, Catherine. So just a reminder to everyone present today, we do have a hashtag. If you'd like to tweet along for the event, we'll just jump back there so we can see it. ARDC Skills 2020. So please join the online conversation around skills acceleration in the national landscape. Thoughts, comments, interesting things to say about today's presentation, very much welcomed on Twitter. And there's a number of other topic hashtags there as well. There's a couple of few quick housekeeping things that I'll run through you through before we jump into the presentations and an acknowledgement of country. Please keep yourself on mute if you're not speaking or asking a question. Each of the speakers will have 15 minutes to do their presentation, followed by five minutes of questions. So please very much encourage you to put questions, comments, thoughts in the chat. If you'd like to ask a specific question at the end of the speaker's presentation, but you want to reserve that for intrigue and excitement of everyone else, please just put your name in the chat and say that you'd like to ask a question and I'll call on you when it gets to the Q&A section. So before we begin and hear from our first speakers, I'd like to acknowledge the traditional custodians of the lands on which we're all meeting, many of us from lots of different places virtually around the country. I personally am joining from the lands of the Wurundjeri people of the Kulin Nation and I pay my respect to Wurundjeri elders past, present and as well to any Aboriginal and Torres Strait Islander people present in this call today. I also acknowledge the ongoing injustices still faced by Aboriginal and Torres Strait Islander people in this country today and note that the lands on which we're all meeting sovereignty has never been ceded. So jumping into our first chat for today, I'm very happy to introduce you all to Kamanthi Parmadabhan and Dr Titus Tang. Kamanthi is a platform leader at Monash University. Titus Tang is a trainer on the Monash data science and AI platform. He leads the effort in the design, development and delivery of data science and AI training courses intended for Monash researchers and students as well as to the wider research community. Titus has a background in deep learning and computer vision and he completed his PhD in computer vision engineering at Monash University. Unfortunately, don't have a slightly longer bio for Kamanthi, but I'm sure you'll be able to hear all about thoughts and research from our first speaker very shortly. So jumping into our first presentation, accelerating skills development in data science and AI at scale. I know I have said my profile through. I'll give a quick introduction, but I think Titus is going to speak about most of the part of the presentation. So good morning, everyone. I'm Kamanthi. I'm not sure if any of you know me, but I lead the data science AI and sensitive research data platforms at Monash University. So we are part of the Monash technology research platforms and work closely with Monash Research Center and massive computer cluster. So the data science and AI platform was established to be a stepping stone for researchers from all the disciplines from arts and law to engineering and medicine. Our goal was to help them apply machine learning and artificial intelligence to their data intensive research. And that itself is a big statement. However big we are as a team, we won't be able to go out and help every researcher at Monash with their data science needs. So upskilling our research community and building a practitioner community with data science skills is key for us to sustainably achieve this. So in this journey, we create and leverage new and existing training capabilities both within and outside Monash University. So as I said, I'm going to give a very brief introduction about the platform before handing it over to Titus. I think time has given a good summary of Titus profile there. So what is this platform? So the platform is basically a handful of ML and AI practitioners, how they're helping out our research isn't a specific research problem. To be collaborate with these research groups, co design and co build some of the niche research techniques, pipelines and workflows. Along with our researchers, we also pilot some of the cutting edge techniques, like federated machine learning to research programs around sensitive data. Our aim is to create a catalog of various data and tools available for our ML community and support our research is to use and apply it to their research. In order to do all this, we do a range of activities to support the support and accelerate the adoption and application of AI and ML, specifically for our nice communities. One of the key things which we have found really useful is the information sessions, bringing some awareness about AI ML, what are the exemplary projects in this space, which other groups nationally and internationally is done, done, and and providing specific training and targeted workshops. These are the key for us to achieve this. That's where most of our conversation starts. From there on, we engage with them on specific research challenges, either through drop in sessions or one on one support, or even directly working with them in their research projects. So the conversation slowly progressed towards building community of practices by bringing together research is applying similar techniques and connecting with expert researchers, specifically from faculty of IT and faculty of engineering. Communities of practices more of a peer support network that we build internally within Monash, but obviously look at expanding it much beyond Monash Partis. And as I mentioned earlier, we pilot some of these techniques like federated machine learning, where data cannot move out of the upper sizes, but model can travel around to learn from different data sets. This is applicable right from hospital data to smart energy data to child protection data, and so on. But within the whole Monash landscape, we're just part of a bigger pie, or actually part of the pie pyramid here. We underpin the Monash Data Futures Institute by bringing user communities and practitioner communities together. We are also underpinned by Monash data fluency, specifically around trainings that David Buonavik and will speak about in the last session today. I'll hand it over to Titus, and probably we'll take the questions at the end. Thanks everyone. Thanks Kamati. Hi everyone. So what I'm going to do over the next few slides is just to give an overview of the activities that we have conducted over the last six months to give you a few of our approach of how we scale up our training and the approaches that we have taken. So as a relatively young platform, we pretty much started from scratch, from zero, and we had to design all our workshop contents from scratch. Over the last six months or so, we have conducted approximately 55 hours of trainings, involving more than 400 people across 20 different research government and commercial organizations. Most of our attendees are from Monash itself, but we do have special events open up to national level attendees, which I will discuss over the next few slides. So our workshops have received pretty good feedback over the recent months, and it seems to be very popular. I don't recall a single event where we have not, where any of these workshops were not fully subscribed, and it is very typical where any one of our events would have a waiting list that is as large as the capacity of the workshop itself. Overall, we have contributed approximately 400 plus hours of workshop content development and delivery time through the lead instructors and the teaching assistants. All of our workshop materials open source and available online through the links provided on this slide, and also available on our website. So part of our training and workshops are funded by the ARDC platform project called Environments to Accelerate Machine Learning Based Discovery. This is a project in that we partner up with the University of Queensland and the Queensland Cyber Infrastructure Foundation, QSIF. So through this partnership, we are conducting a range of activities, as listed on this slide. One of the first things we did was to conduct a combined national survey of the needs and gaps in skills and trainings in the areas of data science and AI, and specifically in deep learning. So I will talk about the results of that survey over the next couple of slides. One of our ongoing activities through this arrangement is to open up seats for sharing between the partner institutions. So for example, Monash would open up and share seats at each of our workshops to QQ and QSIF, while it's versa. And in the short to medium term, we are looking at sharing and up stealing associate instructors and helpers to help scale up the running of our workshops across institutions. We are looking at possibilities of developing combined training materials in order to accelerate the development of new workshops and to eliminate any overlap that we might have across institutions. We intend to publish a repository of open source trainings and materials. In fact, we have already done so, as I mentioned on the previous slide. And we are also planning to embark on training the trainer activities in order to keep our the runnings of our obvious events sustainable in the longer term. Now, so I'm going to provide more of these more details relating to this partnership in a separate learning talk that's happening tomorrow. So if you feel free to hear about the details, please join that session. But I just have to emphasize a point that even though this partnership is currently between Monash, QQ and QSIF, we are certainly open to further collaborations across any institution across Australia. So if you have the same thoughts as us in terms of upscaling, sorry, in terms of upscaling to conduct the the running of these workshops on a national level, please do get in touch with us. Now going back to my point, we've recalled from before that I mentioned that we have attendees from approximately 20 institutions across Australia attending our workshops and many of them attend our events through what we refer to as national workshops here. So the national here just means open to everyone in Australia to attend and the intention for running these workshops is to get a feel for the needs in the community as to what kind of training events will be helpful for researchers across Australia and also to get an idea and a sense of the community that could emerge if we were to continue on these training events in the longer term. So we have planned for three national workshops through the months of September and November. In fact, the last one in this series is running next week. We have so far received pretty good reception. We are always fully booked with a waiting list and many attendees from the first bookshops are returned into subsequent workshops. So we look forward, we definitely look forward to run more of these events next year to build up that national community. Going back to the survey that I mentioned earlier, this survey was conducted between Monash, UQ and QCIF. We have received 161 responses across 18 institutions nationwide and this is just a single snapshot of the various results that we have received and the key question on this slide here is what areas of research in deep learning most interest you in the context of applying deep learning to your research? And to our surprise, semi-supervised and unsupervised learning were at the top of the list in terms of interest from researchers and I guess that speaks to the one of the main challenges in deep learning in terms of getting labeled data for training a deep neural network. So as a platform, our aims are certainly to support researchers in this area and in fact we are in the process of developing a semi-supervised deep learning course targeted at these individuals. So that's a quick summary of the training side of things on the platform. Now going back to the ARDC ML project that I mentioned about before, we are in the process of developing a national web page that brings together various machine learning resources, expertise, training materials and connections for Australian researchers. Komati will be providing more details about this platform in a separate talk so I'll leave that to her but I'll just like to mention just that invite everyone who's interested in in the sharing of events at national level to do visit the website, check it out and we have options on the website in which you could add your own workshops for public viewing. So we have training events that up upscale researchers, we have drop-in sessions that have been created and mentioned so far that allows us that allows researchers to follow up on any questions they might have after the training events. What happens after that? Well we are looking at establishing as a platform various committees of practice across Monash University. These committees of practice serve to allow researchers in a specific niche to come together to share knowledge, to share experience, to talk about what should be done or what should not be done in order to incorporate data science and AI into their workflows. So one of the first committees of practice that we have established is the NLP community of practice at Monash. We started off with a showcase event allowing researchers maybe 10-15 minutes to talk about their topics with the local community and to introduce themselves. Our role as a platform is then to connect these researchers to practitioners and fundamental NLP research experts to be able to leverage of existing expertise and knowledge. Our role as a platform is also to introduce various pre-assisting tools such as COX-TAC that NLP researchers who are new to the field leverage instead of building things from scratch. And we do conduct recurring community meetings similar to the showcase events to encourage community cohesion and to make sure that no one is left behind. On a more regular basis, on a more day-to-day basis, we have slack channels in which researchers could present their impromptu ideas and to start discussions. And as a platform, we also provide very relevant direct hands-on technical support to these researchers perhaps in the form of having an NLP expert or a researcher or a programmer that really assists in a project from perhaps one or two days a week. And last but not least, industry partnerships. So as a platform, we are fully aware that there are various industry partners out there that have a whole host of learning tools, materials and certifications that we could leverage on. So obviously our role as facilitating researchers access to these industry tools and materials to accelerate their research. So by industry partnerships, we are referring to partnerships with the NVIDIA Deep Learning Institute, the Microsoft AI School and AWS Educate, for example. And all of these connections with industry are facilitated through the Monash Data Futures Institute.