Machine learning makes sure that our future seaweed diets are viable and delicious

Seaweed farming is an exciting, burgeoning industry, that creates a viable foodstuff for both animals and humans, ingredients for wastewater treatments, and biomass for fuels, in previously un-farmed parts of the world. It also helps to reduce greenhouse gasses in our atmosphere, and lower ocean acidification. It’s a win on all fronts for a global society that is likely to face increasing impacts of climate change in the coming decades, including food shortages.

But as with all farming, warming seas from climate change and disease in aquaculture farms are a major concern, and in a rush to focus on the fastest growing, highest yield, and most profitable strains of seaweed, there’s a real danger of entire crops being lost to pestilence and environmental changes

Machine learning is playing a key role in helping to avoid that tragedy.

A small team out of MIT are using 2D images taken from a holographic microscope, to extrapolate an understanding of the microbiome of the seaweed farm using machine learning. These advancements in machine learning imaging have been combined with a low-cost communication device which allows the information to be sent to the surface quickly and easily, making it much more affordable for farmers and to gain access to this potentially game changing technology.

It’s hoped that this snapshot of the seaweed’s microbiome will allow farmers and aquaculture enthusiasts to spot disease before it gets out of control, allowing them to quarantine or cull affected stock to prevent it spreading.

It may also inform future years’ seaweed types, encouraging diversification to prevent single-species farming from being devastated by a single pathogen.

The team behind it hopes that this also encourages people to think more about the ecosystems that farming occurs in, but even if it merely helps improve the yields and effectiveness of seaweed farming in the future, the benefits could spread far and wide.