According to the National Oceanic and Atmospheric Administration, aquaculture represents a $1.5 billion annual industry within the United States. Like land-based agriculture, shellfish aquaculture requires healthy seed production to be a sustainable industry to take care of. The production of shellfish larvae (seeds) in aquaculture hatcheries requires close monitoring to trace mortality rates and assess health status from the earliest life stages.
Careful remark is required to influence production planning, determine the impact of naturally occurring harmful bacteria, and ensure sustainable seed production. This is a necessary step for shellfish hatcheries, but is currently a time-consuming manual process vulnerable to human error.
With funding from MIT's Abdul Latif Jameel Water and Food Systems Lab (J-WAFS), MIT Sea Grant is working with Associate Professor Otto Cordero of MIT's Department of Civil and Environmental Engineering, Professor Taskin Padir, and research scientist Mark Zolotas at Northeastern University Institute for Experiential Robotics and others on the Aquaculture Research Corporation (ARC) and the Cape Cod Commercial Fishermen's Alliance to advance technology for the aquaculture industry. Based in Cape Cod, ARC is a number one shellfish hatchery, farm and wholesaler that plays a very important role in providing prime quality shellfish seeds to local and regional growers.
Two MIT students joined the project this semester, working with Robert Vincent, associate director of extension services at MIT Sea Grant, as a part of the Undergraduate Research Opportunities Program (UROP).
First-year student Unyime Usua and second-year student Santiago Borrego use microscopy images of shellfish seeds from ARC to coach machine learning algorithms that help automate the identification and counting process. The resulting easy-to-use image recognition tool is designed to assist aquaculturists distinguish and count healthy, unhealthy and dead shellfish larvae, improving accuracy and reducing effort and time.
Vincent explains that AI is a robust tool for environmental science, enabling researchers, industry and resource managers to handle challenges which have long been a sticking point for accurate data collection, evaluation, forecasting and streamlining processes. “Financial support through programs like J-WAFS allows us to handle these issues head-on,” he says.
ARC faces challenges in manually quantifying larval classes, a very important step within the seed production process. “When the larvae are in the expansion stage, their size is continuously measured and counted,” explains Cheryl James, head of larval/juvenile production at ARC. “This process is critical to advertise optimal growth and strengthen the population.”
The development of an automatic identification and counting system will help improve this step within the production process with time and price benefits. “This is just not a straightforward task,” says Vincent, “but with the guidance of Dr. Zolotas from the Northeastern University Institute for Experiential Robotics and the work of the UROP students, now we have made solid progress.”
The UROP program advantages each researchers and students. Involving MIT UROP students in developing some of these systems provides insights into AI applications they may not have considered, providing opportunities to explore, learn, and apply while helping to unravel real-world challenges.
Borrego saw this project as a chance to use what he learned in Class 6.390 (Introduction to Machine Learning) to a real-world problem. “I began to assume how computers could see images and extract information from them,” he says. “I desired to explore that further.”
Usua decided to pursue the project since it could have a direct impact on the industry. “I'm very serious about seeing how we will use machine learning to make people's lives easier. We are using AI to assist biologists simplify this counting and identification process.” Although Usua was unfamiliar with aquaculture before starting this project, she explains, “When I heard in regards to the hatcheries that Dr. As Vincent told us, it was unlucky that not many individuals knew what was occurring and what problems they were having on the door again.”
On Cape Cod alone, aquaculture costs $18 million annually. But the Massachusetts Division of Marine Fisheries estimates that hatcheries can only meet 70-80 percent of seed needs annually, impacting local producers and the economy. With this project, the partners aim to develop technologies that increase seed production, improve industry capability and help understand and improve the hatchery microbiome.
Borrego explains the initial challenge of working with limited data. “In the start we needed to undergo and label all the info, but this process helped me learn lots.” In true MIT style, he shares his learnings from the project: “Try to work with the info you may have to work with “To benefit from what you're given.” You need to adapt and alter your strategies depending on what you may have.”
Usua describes her experiences through the research process, communication throughout the team and deciding which approaches to take. “Research is a difficult and lengthy process, but there may be much to be gained by learning to search for things yourself and find your personal solutions to problems.”
In addition to increasing seed production and reducing human labor within the hatching process, staff expect this project to contribute cost savings and technology integration to support one of the vital underserved industries within the United States.
Borrego and Usua each plan to proceed their work with MIT Sea Grant for a second semester. Borrego is serious about learning more about how technology may be used to guard the environment and wildlife. Usua says she hopes to explore more projects related to aquaculture. “It looks as if there are countless ways to handle these issues.”
