Teaching and Training Biology Undergraduate Students Computational Skills to Address Marine Ecological Questions

Many of the important issues facing the marine realm and more broadly the natural world cannot be addressed by focusing on one discipline alone. Furthermore, meaningful careers in marine science will increasingly require complex, skilled, non-repetitive tasks and knowledge that cannot be automated away. Training and teaching people to tackle these issues and acquire these skills can be particularly challenging when they have previously limited exposure to marine science, computational skills, and research in general. In this presentation, I describe techniques and approaches to teach and train biology undergraduate students to address marine ecological questions concerning plankton population dynamics using machine learning. This interdisciplinary work involves the synthesis and integration of plankton ecology, computer programming, deep learning, and science communication. The approaches include appropriately motivating marine ecological questions, emphasizing the usefulness and transferability of methodologies, a flexible framework for involving researchers with different backgrounds, and knowledge of tools to teach skills, collaborate, and transfer information about plankton ecology, programming, and deep learning. These techniques and resources will be presented in the context of examining relative changes in marine heterotrophic protists abundances using deep learning. The practices and tools can easily be appropriated and modified for other diverse, interdisciplinary marine science research.