About

Hello! I’m an AI researcher at the Air Force Research Laboratory, working at the frontier of human-machine collaboration. My research focuses on developing adaptive AI systems that can co-evolve with human experts, enhancing decision-making in complex, mission-critical domains.

Research Interests

• Developing transformer-based architectures for real-time adaptation to individual expertise
• Integrating cognitive science principles with advanced machine learning to model expert decision processes
• Investigating attention mechanisms optimized for sequential preference learning and expert behavior modeling

Recent Work

Evaluating Cognitive Flexibility in LLMs

• Adapting neuropsychological assessment paradigms to probe task-switching in transformer models
• Key finding: Performance differential between GPT-4 and other LLMs, with implications for system design

For more up-to-date recent work, please visit my Google Scholar page.

News

[July 2024] Presented our work investigating the cognitive flexibility of LLMs at the International Conference of Machine Learning Workshop on Large Language Models and Cognition in Vienna, Austria. This work highlights the importance of architectural choices in enabling robust task-switching capabilities.

[July 2024] Presented our work on integrating cognitive models of human perception and vision transformers for adversarial robustness at the International Conference of Cognitive Modeling at Tilburg University, the Netherlands. This research showcases the potential of aligning AI systems with human cognitive processes to enhance robustness.

[December 2023] Ryan McCoppin presented our work on evaluating adversarial robustness at the IEEE International Conference on Machine Learning and Applications (ICMLA). This research contributes to the development of more secure and reliable AI systems.

[August 2023] For the second year in a row, the 711th Human Performance Wing selected me to mentor an intern over the summer. Goonmeet Bajaj, a Computer Science and Engineering PhD candidate at the Ohio State University, worked with us on addressing knowledge gaps in visual question answering systems. She also worked with us on leveraging LLMs for cognitive architectures and presented this work at the AAAI Fall Symposium Series on the Integration of Cognitive Architectures and Generative Models. These projects exemplify our commitment to advancing human-machine collaboration across diverse AI domains.

[April 2023] Received a DoD SMART Retention Scholarship. I will be starting fall 2023 at the University of Wisconsin - Madison in the Computer Science PhD program. This opportunity will allow me to deepen my research into adaptive AI systems in close collaboration with leading academic experts.

[August 2022] The 711th Human Performance Wing awarded a summer internship to Dante Miller, a Computer Science PhD student at Rice University, who I mentored on a research project focused on Case-Based Reasoning for Explainable AI in a Clinical Domain. This project aligns with our broader goal of developing AI systems that can effectively support human decision-making in high-stakes environments.

[July 2022] Attended ICML 2022 in Baltimore and presented a blue sky idea on human-in-the-loop implications for robust ML. This work sets the stage for our ongoing research into human-AI co-evolution and adaptation.

Through this work, I aim to advance the frontier of adaptive, cognitively-aligned AI systems - enabling more effective human-machine collaboration across a range of complex, real-world domains.