Algorithm to help understand wild fishes
Researchers have developed an algorithm to help determine the growth and wellbeing of wild fishes. The algorithm provides biologists with a long-needed tool to very accurately measure the effects of environmental impacts and climate change on the growth of wild fishes.
Algorithms can be used for much more than just controlling robots or changing your feed on social media. By pairing a biologist and a researcher in artificial intelligence, it has proved possible to develop an algorithm to determine the growth and wellbeing of fishes in the oceans.
- Using machine learning, we have found that if we provide the machine with data on the weight of the fish along with two biochemical factors about the fish’s digestive system and protein deposition, the algorithm can accurately predict the growth rate, explains Poramate Manoonpong.
He is Associate Professor at SDU Embodied Systems for Robotics and Learning, University of Southern Denmark.
Much-appreciated innovation
Until now, biologists have had no accurate way of estimating the growth and wellbeing of the wild fishes, so the algorithm is a much-appreciated innovation.
- Our work will be a major help to minimize uncertainty in wild stock assessment process. It provides us with a unique look into the effects of environmental impacts and climate change on the growth and wellbeing of aquatic living resources through the key biological factors representing the actual responses of the animals themselves, says Dr. Krisna Rungruangsak-Torrissen, a nutritional biochemist from the Institute of Marine Research, Norway.
A world of numbers
Poramate Manoonpong has gotten himself deep into artificial neural networks for building robot brains.
In the branch of bio-inspired robotics and neurorobotics, he has used the brains and nervous systems of animals as inspiration for the design of the robot brains.
Krisna Rungruangsak-Torrissen has spent over 30 years building a detailed database of the salmonids, and a conversation between the two has created collaboration between the two different fields where she provided the ideas, the collected data and knowledge with Poramate Manoonpong, who then began filling it all into mathematical models.
”There is no question that it is a personal motivation to contribute to the environmental agenda
- When I work, I see numbers. I don’t really think about what the number is; e.g. if it is from fish. I only focus on the numbers and the math to avoid limiting myself, but there is no question that it is a personal motivation to contribute to the environmental agenda, says Poramate Manoonpong, who developed the algorithm in his spare time.
- Artificial neural networks have been used to predict the weather and make robots move, but this is the first time we have seen that they can be used to predict the growth of animals.
Poramate Manoonpong and Krisna Rungruangsak-Torrissen are now in contact with researchers in Thailand, who wish to co-operate with them on the algorithm to track the fish and shrimp stocks.
- Read the research article: Neural computational model GrowthEstimate: A model for studying living resources through digestive efficiency
- Read more about SDU Embodied Systems for Robotics and Learning
Meet the researcher
Poramate Manoonpong is an Associate Professor at SDU Embodied Systems for Robotics and Learning. He is doing research in the field between bioinspired robotics and neurorobotics. The brains and nervous systems of animals have inspired him to design artificial brains.