 So, let's move on and welcome our next presenter, which is Sherry Merlin, CEO of kawalai.co, talking to trace AI on traceability access for consumers and exports. Welcome again, Sherry. Good day. I'm Sherry Merlin, founder and lead innovator of Kowil AI from the Philippines and in the Southeast Asian region. We are Kowil AI, an industry agnostic artificial intelligence solution with customized computer vision tools for machine learning models that can be integrated to mobile and web applications. Our mission is to help humanities understand artificial intelligence. Our vision is to provide AI solution to preserve biodiversity, livelihood and environment, reliable and unbiased data analytics on demand. Our solutions addresses sustainable development goals number nine, innovation and infrastructure through our smart products. SDG number 12, Responsible Consumption and Production and SDG number 14, Life Below Water. Our core team has more than 20 years of combined experience in building AI products and services from research to commercial application, providing state-of-the-art AI tools. We are the frontrunner in computer vision AI in the Philippines. The technical team is rooted in AI academic research and with experience in cybersecurity and autonomous systems. The main problem that we are trying to solve is the illegal, unreported and unregulated phishing in the Philippines. Overcoming the pen and paper documentation method that still exists in countries like ours. The current pandemic has accelerated the need for the digital economy, especially in the food supply chain, to keep up with the increasing consumption of food, including seafoods. Digital illiteracy is the strongest barrier to technology adoption. The Philippines has one of the longest coastlines in the world, yet our fishermen remains the poorest of the poor due to the lack of resources and knowledge for documentation. Our fishermen sell cheaply or resort to illegal, unreported and unregulated phishing for fast money. Our solution, currently in beta version, is trace AI, or traceability access for consumers and exporters powered by artificial intelligence, is an automated electronic catch documentation and traceability mobile application, which was pilot tested with WWF tuna sustainability program in Mamborao and Sablayan, Occidental, Mindoro last 2019. With initial registered 53 fishermen, 250 images collected, current accuracy at 80% and currently identifies yellow fin tuna species. How does our mobile application works? Trace AI on its beta version has leverage AI and machine learning to reduce the friction in data gathering for fishermen and eliminate fraud. The mobile application works offline. The scan, snap and send is the official method to record the catch data. This uses AI for fish identification, estimates the length and weight of the fish, specifically tuna, auto catch recording with geolocation and vessel information. Upon arrival to the landing port, the information is then uploaded to the cloud via mobile data or Wi-Fi. Due to the pandemic, we have geared enhancement of the system that will include additional seafood species for identification and traceability. It will include a seamless and interoperability system that can be available to government and private stakeholders via web application and cloud system. Lastly, is the marketplace and logistics integration to complete the catch the plate journey via QR code labels. If we look at the business aspect of the seafood traceability, I do believe that it is worth the hardship for bringing it to the market. In 2020, the global market for seafood traceability is $500 billion and in the Philippines, it's currently $7.2 billion of market opportunity. This solution is very much scalable as it utilizes AI and data analytics as the future of traceability. We are here to share our knowledge and solutions and to be able to learn from others their best practices in AI data training and implementation with possible collaboration for this initiative. We look forward to number one, the creation of database system that can be accessed openly for data training to help increase AI accuracy of tools. Number two is to explore simplified AI applications that can be implemented using existing gadgets and mobile devices, making it cheaper and easy to implement. And third is the AI integrated research collaboration. Development of AI tools is tedious and expensive and research funding for AI initiatives should be available to non-academic space. To have a more granular understanding of grassroots application. For the adoption of the digital blue planet and to provide a thriving ecosystem, we need to involve the coastal communities in this initiative. The goal of Kawil AI is to leverage the data collected by the fishermen for them to access financial support like personal loans and grants. Second is the access to the tools for our fisher folks to have access to hardware to be able to utilize the technology for their own good. Third is to incentivize the fishermen to have the value of the data they contributed for the AI database and development. In summary, we have a built customizable mobile application powered by artificial intelligence for seafood traceability, which currently identifies tuna species with 80% accuracy using a mobile phone application with ongoing enhancement for additional seafood species application and marketplace platform. We are Kawil AI Solutions bringing AI technology from lab to life. We want to hear your thoughts and kindly reach out to us through our digital directory. Thank you and have a great day. Thank you very much for sharing the work of your team, Sherry. It's already quite apparent just how global the initiatives are just in the first three presentations we've had and the nascent knowledge that's being built up. I've got a range of questions for you, but I'll try and package them and you can choose which ones you like. I'm interested largely in the challenges you've overcome that you've encountered and overcome and especially about how you select the metrics that you think are going to be most important to hone in on. For example, what species you choose and what species you'll choose next. And I'm just wondering how you foresee the challenges that you've had for others coming into this situation and how there might be a requirement for for international co operations around the standards we all use. So when it does come to cross collaborating we find this a lot easier. Thank you. Thank you, Kim, for that question. So, you know, our matrix came from also from the community. So we here in the Philippines, Tuna is like number one export species for seed food. But as of now, we fall short for the supply because of the illegal activities because of fast money as you know, but with the support also of the international institution, we kind of pin down what are the things that they need or the constraint that they're having why they don't do traceability or documentation. So number one is the understanding of the value of really documenting and selling the product on a fair price. As you know, even if it is fast money, it's very easy to spend. It doesn't have value and you know, you can repeat that sales again. But for our fishermen, we want them to really find out the value of the sustainability side and like complying with the necessary information and looking forward to the future to be included to that, you know, market of seafood and supply chain. So that's how we do our metrics. And I think, yes, it can be replicated in other countries. And we're looking at one collaborating with other Asian countries for now and hopefully in other tropical countries that has this specific requirement. And to add to that also our metrics are patterned with the European compliance of export market for tuna. So the metrics that we're collecting through our mobile application is used to that requirement. So, yes. Sorry, if Matt give me one more go up. I'm just very interested. We've also developed a an app and looked at developing algorithms to look at a range of sharp speaking and we did have challenges in trying to work out how to get the best accuracy out of the algorithms and you mentioned that you get 80% And I'm sure that would improve all the time. I'm just wondering was most of the error margin coming from the way images were taken the number of images in your training data sets give us some ideas on on on the the process of how you've done this and where you've managed to overcome anything or some ideas on what's helping you improve your algorithms for species identification. Yes, the most important the most I think challenge that we have encountered is how to take the picture correctly so that the algorithm can accept it for the 250 recorded. It's a it's accepted images out of like, almost 1000 images that we have collected on several occasions so out of that and our algorithm, having increased the accuracy for for for identification. One is we need to have a good way to take the pictures, and that is why we need coordination with the fishermen and what we call in the locals in new veritors. So they are the ones teaching the fishermen how to take pictures. And that's why I said we need to communicate with the community so we can have better data. So that's number one requirement in our training and for now we are really focusing on increasing and being agnostic in our identification. That's why the enhancement of our app is focused on different species that are highly consumable on international market. Thank you. Hi Terry thanks for amazing work you're doing that. Hi, I could talk for hours with you about regulated illegal fishing in the Philippines and what goes on that. But what I was interested in is how the fishes interact with you and how you're building a community of practice. I have a lot of social incentives in there as well I worked on an active unit that was called environmental witness. And in the Philippines, everybody talks to one another, especially about fishing. And that works really well. So if you have a social aspect to your to your work as well, you know, developing people reporting to one another about how well it's going or if they've seen some of naughty or anything like that. Yes, that's how we started with Fisher folks communities. We do interact with them. We build our network and we tell them, you know, the idea came from, I was doing before marine coral mapping. So we rent boats. Sorry for the short story. We rent boats and we rented the Fisherman's boat. So while on the sea, he was watching like online shows. So we said, that can be utilized, you know, accessing that data or the information online. Having seen that Fisherman accessing online gave us an idea to create this app. So we tend to really tell them you want to know more. So for now, it's more on educational aspect of social building, but we're looking at really incentivizing or having partnering with stakeholders or government to really support that, you know, initiative in documentation. So that's the overall goal in the social side of like a wheeled with our trace AI application. So yeah, we have seen that they have access in some areas. So we need to utilize it. And right now with current situation of the pandemic, having more access to, you know, internet and Wi-Fi, even here in the Philippines, they're increasing the internet connection. So we want to utilize it more on the community side, especially on the fishing community. I'm plastic cherry. Yeah, there's a real role for social media at this time, isn't there? Yes, yes. There's a lot of people posting their their catch online on Facebook and other things. Exactly. It works really well. Thank you so much. Thank you. Thank you. Thank you, everyone. Yeah. Please do connect. Thank you very much.