Machine learning and artificial intelligence, including deep learning methods, have become standard tools in many fields. The technology can be applied directly to several data analysis problems and have been shown to solve highly complex data analysis challenges.
Advances in methods, sensor technologies, and autonomous platforms allow us to scale up data collection programmes, but as long as the data processing requires manual scrutiny and does not scale accordingly the benefit from increased data volumes is not realized. Machine learning offers ways to automate data analysis and address the analysis bottleneck. It may also prove useful in detecting patterns and shifts in highly multidimensional systems such as dynamic ecosystems and socio-ecological networks, which may be difficult to analyse with more traditional statistical methods.
Theme session C will begin with invited talks. There is a wealth of information of variable accuracy related to machine learning and AI, so the objective of the talks will be to demystify the methodology and offer a gentle introduction to the topic. The talks will also outline typical challenges for implementation, including data preparation and validation frameworks.
The second part of the theme session seeks contributions from the community, where machine learning and AI have been applied to problems within marine science. We encourage talks that explain why the methodology was chosen, what alternatives were considered, and what kind of problems were encountered as well as how they were solved. We welcome both applications of traditional ML algorithms and novel state-of-the-art methods.We aim to cover a broad variety of topics from across the marine sciences.
Theme session C concludes with a discussion of the possibilities and challenges of using machine learning within marine science, and the objective is to provide recommendations on how to move forward with machine learning within the community.