Machine Learning in Ocean Pollution and Beyond

Machine learning has had its breakthrough as a field predominantly after the 1990's. Partly because the computing power needed for machine learning was not available until after. As a result of the novelty of machine learning, new applications are being found for it every year. Its ability to compute numerous factors and provide reliable estimates makes it applicable anywhere big data is involved. I believe that its applications can also be extended to aquatic big data. For example, "Satellites And Machine Learning Are Being Used To Detect Plastic In The Ocean":

Recently, I read an interesting article by the NY Times about how machine learning is being used to find humpback whale sounds. Researchers collected 180,000 hours of recordings, and wanted to filter through other noises in order to determine when and where the humpbacks were singing. Due to the sheer size of the data, manually filtering through it would be undoubtedly tedious, and it would lead to errors. Because of this, the researchers approached a Google engineering team to create a machine learning model for it. Within just 9 months, Google had a framework for the humpback whale sounds. After finishing the model, they even adapted it to use in Canada's fisheries.

I feel like this is a great example on how machine learning can be used to filter through big data and create meaningful models from it. The EPA, for example, is now implementing sensors to detect trash in the ocean. Machine learning can, and likely will be used to filter through the chemical analysis of trash and identify different kinds of plastic. I feel like machine learning will also be the next big thing in predicting ocean acidification. pH sensors in bodies of water all around the world produce monumental amounts of data that is conventionally incomprehensible at a high level of abstraction. However, machine learning would be able to somewhat accurately factor in the aquatic big data produced by these sensors and output a model for predicting the decrease in pH. If scientists can correlate the models with CO2 emissions, we can predict the measures we need to take in order to stabilize aquatic ecosystems. I strongly believe that the implementation of machine learning models would be a great step in the right direction, and would help us prevent, rather than cure, the decline of aquatic ecosystems worldwide.

"Machine learning is helping to track the giant larvacean, whose mucus houses trap carbon dioxide, sending it to the bottom of the ocean."

Read More + Source!