About 20 students will work closely with AI, cybersecurity, and marine science partners to test and prototype techniques to identify and automatically mitigate data poisoning in oceanic computer vision. In collaboration with VIMS, NOAA, and the NSF, this project will focus on securing AI systems designed to be leveraged by the fishing industry, focused on benthic biomass estimation. The student fellows will be working closely with them in an experiential learning context to investigate how adversaries might leverage data poisoning to disrupt species detection and spatial ecological inference in image-graph fused AI systems. Students will design, test, and defend deep learning pipelines that combine HabCam imagery with graph neural networks to model marine species distribution, which is critical to fisheries monitoring and management in ecologically sensitive areas like the Chesapeake Bay. The program will continue to expand Virginia's cybersecurity talent pipeline while advancing foundational research in adversarial machine learning, contributing to national security, and supporting the commonwealth's long-term economic resilience.
We are soliciting applications to the 2025-2026 CCI Fellowship program until November 15th, 2025. Award recipients will be notified by early December. Students selected for engagement will become eligible for monetary support from a prize pool, pending participation and success in the program. The top performing team in each of the three focal areas will receive a prize.
In addition to providing an opportunity to get real-world research experience during your time as a fellow, students that are selected will have access to the following resources and opportunities: