MIT Researcher’s Machine Learning Study Can Save Seaweed

Seaweed is very popular in East Asian cuisines, and it has enormous promise as a long-term food supply for the world’s rising population. Seaweed, in addition to its nutritional value, protects the environment from a variety of hazards. It aids in the battle against climate change by absorbing extra carbon dioxide in the atmosphere and fertilizer run-off, keeping coasts clean.

Seaweed, like so many other marine species, is endangered by the exact thing it serves to mitigate: climate change. Warm temperatures and little sunshine increase the development of dangerous germs such as ice disease bacteria. Due to unregulated bacterial proliferation, entire seaweed farms are destroyed in days.

An MIT researcher, Charlene Xia, and her professor and other researchers propose a study to forecast and control disease transmission in aquaculture. The team concentrated their efforts on seaweed farms in particular. They used the microbes found in these seaweed farms to predict any harm to the seaweed to tackle this problem.

In computing, the team mixes ancient and contemporary technologies. They take a two-dimensional picture using a holographic submersible digital microscope. They then utilize a neural network, a machine learning system, to turn the 2D picture into a 3D representation of the microbiome.

The program may be installed on a Raspberry Pi and used with a holographic microscope. Xia used her master’s degree research to figure out how to transmit this data to the study team. She concentrated on constructing miniature underwater communication devices that could convey data about the ocean to researchers under the supervision of Professors in the Media Lab. These gadgets are meant to cost less than $100 rather than the customary $4,000, lowering the financial barrier for anyone interested in unraveling the myriad mysteries of our seas. Machine learning techniques can be utilized to send data about the ocean environment via communication devices.

The research team intends to build a low-cost, real-time monitoring system deployed to large seaweed farms by integrating these low-cost communication devices with microscope pictures and machine learning. It’s almost like having an ‘internet of things’ underwater. They are working on an underwater camera system and wireless communications that will allow her to get data while sitting on dry land.

Using this microbiome data, the research team can predict when a disease is likely to occur, putting seaweed at risk before it’s too late.

Source: https://news.mit.edu/2021/saving-seaweed-machine-learning-1022

Suggested