 This presentation is by Chono Abeledo talking about CrabTech and this is a very exciting example of the kind of work we're doing. Another really good news story about AI solutions that have moved from the research phase to the practical phase and really helping people on the water. Hello everyone, I am Chono Camil Vincicruz Abeledo, the project leader of the CrabTech team from the De La Salle University from the Philippines and I'm here to talk about some of the technologies we've created to help improve Philippine aquaculture. Our work often starts with a problem with some of them rooted in the difficulties in identifying an organism. Sometimes we are led to work on strange new characteristics or in figuring out the connectedness of things. Most of our work involves the extraction of DNA or RNA from target organisms that we work with. And we use the unique sequences found in these molecules to find a match in existing databases. We also make use of the sizes of these unique sequences to find organisms with similar traits. Using these techniques and with funding from the Department of Science and Technologies, Philippine Council for Agriculture, Aquatic and Natural Resources Research and Development, we have helped various fishing communities. We have helped monitor and confirm the biodiversity of seaweeds in the Verde Island Passage, which is the center of the center of marine biodiversity. We've also found previously unseen connections across the different mangrove crab populations in the Philippines that is now being used to improve management practices of traded crablets. We are also trying to find possible genetic roots for delayed maturation in female crabs that make them appear like an intermediate female type. And with a combination of image analysis, DNA bar coding, and machine learning programming, we have created the free mobile application called Crabbyfire that helps identify to species juvenile mangrove crabs. Crabbyfire started out as a problem rooted in biodiversity and species identification. Because you see, in the Philippines, we have three of the four species of mangrove crabs under the genus Cila. Of the three, one species is preferred because of its faster growth rate and larger sizes upon maturity. This species is Cila serata, which is locally known as bullic or the king mangrove crab. Differentiating the species at adult stages is straightforward, but at least three sets of morphological markers that are very obvious. But in the case of juveniles, there is no obvious physical feature that has been tested and has proven reliable in differentiating the species. This is a problem because the local mangrove crab industry that is now producing more than 18,000 metric tons per year and is second only to China in terms of production, still uses captured juveniles to stop grow-out ponds. And since the ecological ranges of the three species overlap in the Philippines, any captured batch is a mix of the three species. To develop Crabbyfire, we first made use of morphometrics and image analysis to look for physical features that would cluster individuals of the same species. We then confirmed if the patterns we were seeing were equivalent to species groups using DNA barcoding technology. From the data gathered using this technique, the Crabbyfire app was developed using a convolutional neural network that feeds off additional crab images that we have correctly identified. The application, which uses the carapace shape, specifically the frontal lobe spines, can now identify the species of a single crablet in less than a second using an entry-level camera mobile phone. Its current accuracy is at 92.4%, but we are still hoping to increase this by adding more images to our database. Since its deployment last August 2019, we have helped fishermen ensure that they are capturing or purchasing the preferred species. The pipeline that we have created for Crabbyfire can be applied to a huge variety of organisms, and soon enough we can see applications like Crabbyfire used from the pond to the farm, from the forest to the ocean. At the moment, Crabbyfire is identifying juveniles one individual at a time, but we are also hoping to create a sorting device to help improve the efficiency of our mangrove crab ponds. Currently, our fishery-sort crab lets to size manually, but with this device, we are hoping to cut the processing time to a third, increase its accuracy, put in the option of sorting to species, and lower mortality in the process. Currently, we are also developing the Aline-Mango algorithm. It's like a dating app for crablets and potential grow-out ponds. We are looking at information from existing databases for temperature, salinity, and pH of Philippine coasts. And we will match this with these conditions in fish ponds all over the Philippines to help our fishers determine the best geographic source of crablets are for their fish ponds. And that's what we do as practical geneticists in the Crab Tech team. If you have any comments, questions, or suggestions, please don't hesitate to reach out to us in the contact information that we will show on screen. Thank you so much for this opportunity to share our technologies, and we are looking forward to collaborating with you. Thank you very much, Johnna. Wow, what a complex story, beautifully told. You're a real communicator. Not only are you bringing genetics, you're bringing AI together with the end user all the way to a fantastic story. I wonder if you could share with us a little bit about some of the stories you have about making sure the products that you're developing were accepted by the young farmers and people. How did you, for example, when you were trying to build the user windows and so on, how did you be able to understand what was suitable and what wasn't? Some of the story of how you managed to deliver this all the way to the endpoint of the people you get on the ground. Thank you for that question, Kim. I feel a huge responsibility for carrying this story because it's not just my story. It's not just my team's story, but also the story of all the fishermen who were able to help us with this project. I think part of our story is very similar to what Alberto mentioned earlier. When it comes to developing technologies, it has to start with the people. And in the case of crabby fire, it started out with the fishermen who were helping me finish my dissertation in college. I was trying to do population genetics, which is something so abstract compared to what crabby fire is. And a lot of the fishermen were asking me, Ma'am, why are you looking at that problem when we actually have a bigger problem at hand? So they started showing me the crablets that they were purchasing and capturing and their problem for species identification. So crabby fire actually started out as a promise to these people. How can I help them after all the effort that they did in helping me, a young student at that time, complete my degree? So it started out with me asking questions from these people, finding out their needs, finding out their traditional knowledge, which we are continuing to this day. And once we created a functional prototype, we knew that the first thing that we had to do was to go out there outside of the lab to the fish ponds to the ocean and start talking about what it needs. So currently we're trying to figure out a way to change the language of the application so that the names of the species can change from scientific names to all the different local names in the communities that we are helping. And apart from that, even the Sorter, the classifying device that we want to make is something that our fishermen are asking us to do. So a lot of the things that we're doing really is shaped by what is on the field rather than just a random idea or interest that we have. So thank you. Thank you, Shona. I'm going to pass you over to Matt. I've got another 50 questions, but we're trying to get through. So Matt, can you go ahead, please? Thanks so much, Shona. Thanks so much, Shona. Salam at the presentation. One of the, I could also just talk for hours about this. One of the other, I think you've highlighted the key risks are real incentive for a reason to use technology, aren't they? Of course, one of the big problems for people in the crab industry in the Philippines seems to be happening increasingly more frequently over the past 20 years is super typhoons. And I'm just wondering, you know, we saw some terrible damage this year to the crab ponds. We're not integrating any kind of sort of Python alert and offering guidance to stakeholders to protect their stock, because the damage this year has been appalling, actually. And I've seen very little done so far to address this actually. It's disappointing. But it's a very important high quality value sector. I just wondered if you thought about integrating anything to do with that. And that is a very, very good idea, which we might actually start incorporating to our new program called Alin Mango. So Alin Mango is actually a fusion of two Tagalog or Filipino words, Alin, which stands for which, which type, and Mango, the last portion of the Filipino word for mangrove crab, which is Alin Mango. At this point, I would also like to shout out to one of my students, Kenneth Solis, because this is his master's thesis. And he's part of the audience who's watching right now. But anyway, in Alin Mango, what we're trying to do is match the environmental conditions of the crablets to be fish ponds that are going to grow them. And at this point, what our primary motivation was the fact that climate change is also changing the adaptation of our crabs in the different locations of the Philippines. And it would be best if crabs from a particular environment will be grown to the most similar environment that they could possibly be in in order for us to maximize their growth. And since you mentioned the possibility of typhoons in the Philippines, sadly, is part of the typhoon belt. We could incorporate an early warning system to this. So at least we can forward our mangrove crab fishers and growers to either start harvesting early, or maybe even delay the grow out of their crabs if a new typhoon is coming in. I guess at its core, this is also part of our eventual dream to make a system where we connect our fishers to the marketplace, where in, for example, they can use the same application to find the best source of crablets for using their ponds, project how many crablets they need to buy in order to reach a particular quota or profit margin that they're aiming for. And then at the same time use the same application to sell the crablets to restaurants, markets and supermarkets in the Philippines, or maybe even beyond. So that's that's the dream that we have for the crabby fire system, the crabbeck system. And I'm pretty confident that the same thing has been done or can be done in other fishing, fisheries commodity in the Philippines. So thank you. Just one more, Shana, sorry, we're going to ask you one more question. Anton's going to come with this question. Thank you, Anton. Go ahead, Anton. Sorry, my, my zoom disappeared. I could not find the new button. No, no, great work. Very, very well done. But so how can other countries team up with you and learn from you and how much would this would be replicated in another country with a similar value chain, how much time would it require for students and researchers to collaborate to build a similar project. Wow. That that's a very intimidating question right there. It took us about, I think close to four years to make the first prototype of crabby fire and we are now at our seventh year building our database. Lucky enough, our springboard is population genetics. So we are sort of familiar about the connectivity, how the different populations of mangrove crabs interact, not in the Philippines, but across the Indo Pacific, because it's based on gene sequences that are freely available in an online database. So we know that there are interactions across the different populations, and yet at the same time, we also know that different populations with different environments gain different kinds of adaptations. It would be nice if we can collaborate with locations in Australia in China, and in Southeast Asia that have wild populations of mangrove crabs, because we would like to know how different your juveniles look like from ours. And this can open a portal for international trade, not just at the adult stage, but perhaps even at younger developmental stages to help us improve selective breeding practices. It can also open the pathway for a network, a system of managing this organism because at the end of it, the crabs growing here in the Philippines are not necessarily just Filipinos. The larval stage of the organism is traveling all over the world because they are buoyed by the different flows and ebbs of the water all over the world. So it is a shared responsibility, not just the Philippines, but of other countries. Now in terms of timeframe, I think because we already have the pipeline available, this would be much, much, we can do it much at a much faster rate. And if any are interested in doing this in other organisms, we actually have, we have workshops on the ready. So all you need to do is contact us, and we can set up a virtual workshop for the meantime about how we do the things that we do. But hopefully someday once we're all vaccinated and safe, we can travel over because we do love traveling, and so that we can teach you firsthand about how we create a crabifier and how you can do the same with another organism. Thank you very much, John. And to Kenneth, good luck with your masters and well done for getting such an inspiring supervisor. And just as we kind of hope we're getting questions on your talk from one of the previous presenters who's working in India and as we can see, you know, the efforts and the imagination is, is worldwide. We have people in this presentation from Hawaii, from Africa, from Japan, Holland, and now from the Philippines and if we can start to link these groups to share their knowledge that that would be the successful outcome for this type of event.