A lot has happened during the last year for the victors of the FishAI competition. Since being announced the winners of the FishAI: Sustainable Commercial Fishing Dataset Competition at the Nordic AI Meet in 2022, the team – formerly known as Team Poseidon – has successfully launched a startup commercializing the powerful decision making tool they developed during the competition.
Since launch, Catchwise.no has had over 750 users on their platform representing over 70 shipping companies and are currently working closely with a few selected vessels and fisheries to further develop their product. The FishAI competition focused mainly on reducing Co2 emissions from domestic fishing by creating a user-friendly tool that fishermen easily could employ to reduce time spent searching for fish, reducing unnecessary travel by predicting coordinates to optimize the catch of a specific fish species and creating a sustainable fishing plan that suggests a specific route based on reports. The competition relied on catch data published by the Norwegian Directorate of Fisheries (catch note dataset), temperature, salinity and moon phase datasets.
The tool created by the Catchwise team has shown great potential in ensuring a more environmentally minded fishing industry by allowing fishermen to use data to predict the coordinates of a specific fish. We asked the team about their journey from winning the competition to commercializing their product in Norway.
“The last year has been a ride” – says the team behind Catchwise. “Last summer, NORA’s FishAI competition caught our attention. Using data science and technology to help Norway’s second largest industry, fisheries, to waste less time at sea just sounded interesting. And to help them save 5 and sometimes 18 thousands of liters of diesel per day per boat? Super interesting!”.
However, turning a bright idea into a commercial success is no easy feat. “Winning the competition really tipped us across – we had an interesting and valuable problem, we had a team of interested entrepreneurs and friends, and we had the motivation to build something within the industry”.
One year residence with NCE Seafood and mentorship with support in commercialization was included as part of the prize given to the winning team of the competition. Catchwise.no took advantage of this opportunity and NCE Seafood in turn put the team in touch with some of the key innovators and actors in the industry. “Catchwise would not have been possible without the opportunity to rely and stand on the shoulders of the giants before us. We are also lucky enough to have been able to establish connections with some of the best fisheries in the industry with one goal in mind: to improve efficiency and sustainability for fisheries by leveraging data and artificial intelligence. The most fun has been to travel to the users, going on boats and getting to see our solution in action”.
When asked which advice Catchwise would give future novice AI entrepreneurs, the team replied “Don’t spend too much time building the perfect product, but rather build towards a MLP (Minimum Lovable Product) and get it to customers as early as possible so that you can continuously iterate your product. Furthermore – often a nice presentation of a simple algorithm is much better than a very complex algorithm that is hard to understand for a non-techy user. Learning the domain and the intricacies of the problem you are trying to solve is important and easier with simple algorithms”.
Research and domain knowledge has been important to Catchwise in grounding their project, which has played an important role in their success. The research the team themselves have conducted throughout the competition in the process of launching their startup has been essential and will continue to be useful for them – and others.
“Research has played a large role for our startup, since a lot of what we do requires domain expertise. We have read a plethora of articles with topics ranging from the spawning patterns of mackerel to the latest trends within spatio-temporal forecasting. Building a good data foundation and being pragmatic has also done us favors. We have spent a lot of time finding, putting together and cleaning good data sources, which we thank ourselves for every day. And for pragmatism; trim your problems down as much as possible in the beginning, rather than solving everything at once. For us, that was picking one species of fish to start with; a smaller dataset and a simpler algorithm. By using this method, each iteration when building on the product went a lot faster and we learned along the way. We have 20 years of catch data for over 100 fish species, matched with salt levels, temperatures, wind, moon phase and more data to come. We think future captains, research scientists, students and people in other industries operating in the ocean can learn a lot by looking at this data and conducting analyses, just as we have done ourselves”.
Participation in the FishAI competition inspired the idea and product behind Catchwise. When asked if such competitions can help stir innovation in Norway, the team provided a resounding yes! “The competition inspired us! So many engineers and academics sit and wait for a good business case or valuable problem to help with, and using these competitions to highlight them is gold. For us, this was the perfect nudge for building something that we believe in.”
Want to collaborate with Catchwise.no? The team is highly motivated to continue their collaboration with academia and researchers. The team has built (and are building) a solid data foundation to help identify fish and other types of environmental behavior in our oceans, which could be the basis for many interesting research projects. “For ourselves, we would love to keep working on our predictions to help the environment and fishery. A lot of this work is reminiscent of a research project, so contributions from academia in this field would benefit the scientific community as well as our product”.
The Catchwise team has great ambitions for the future and when asked what role AI can play in succeeding with innovation in the Norwegian fishing industry, the team replied:
“Leading a large boat, whether it is 15m or 100m through the devastating oceans and to capture wild fish in massive quantities is no easy task. We believe that the captains and fisheries we serve are amazing entrepreneurs. Modern technology gives an opportunity for those who are curious to become even better with data, and that's where we want to bridge the gap. We believe that Norway can continue to be a leader in fishing innovation and sustainability, for the rest of the world to learn from”
We congratulate Team Poseidon for their excellent work during the competition, and salute their move from idea to market!
If anyone is interested in getting involved with Catchwise, the team encourages people to reach out through Heisann@catchwise.no, or visit their website, catchwise.no for more information.
For more information about the FishAI competition, download and read the FishAI article published in the Nordic Machine Intelligence Journal.
NORA.ai will continue to host dataset competitions, stay tuned on our webpage and competition portal for more information!