What can we do with the immense amount of data that have been and continue to be collected in aquatic environments? Extracting useful information from “Big Data” may seem daunting, but we explore machine learning and artificial intelligence techniques to process and interpret large data sets efficiently and effectively for fisheries and ecosystem science.
Modern day instruments in marine science can collect data 24/7, 365 days a year for multiple years at impressive spatial and temporal resolutions and coverage, and are rapidly adopted by marine research institutions around the world. The data include acoustic, optic, biological, and environmental measurements that provide unprecedented observations of the physical, biological, and geological environment, and are potentially a boon to ecosystem and stock assessment scientists. However, this abundance of data comes with a suite of challenges, including storage, access and discoverability, processing and analysis, and interpretation, which need to be addressed before these “Big Data” are ultimately useful in a biologically and ecologically meaningful context.
This theme session explores techniques to ingest large data sets and extract useful information efficiently and effectively. New methods to integrate potentially multi-modal data streams can lead to new insights over scales from individuals, stocks, populations and ecosystems and can provide a holistic view of the environment. We welcome examples of these methods to highlight recent innovations in data processing and analytics, in particular at the domain intersection of machine learning and artificial intelligence, and aim to provide avenues for discussion and facilitate interaction among those who deal with “all things data”.
