Research Fellow Candidate Presentations

Please hold this time for presentations from our 2025-2026 Research Fellow candidates.

2:30pm-3:15pm: Richard Hakim – Harvard University

Title: Neural representations of motor patterns and new approaches to study them.

Summary: Motor patterning is a critical function in animals, agents, and robots. We developed novel approaches to study how motor patterning is represented in the brain and evolves over learning. We find that primary motor cortical activity reflects the oscillatory envelope of facial movement, contrastive learning achieves human-level performance in tracking neurons over sessions, and neural activity patterns present during naturalistic behavior are useful for defining BMI decoders.

3:15pm-4:00pm: Ruojin Cai – Cornell University

Title: Pushing the Boundaries of 3D Spatial Understanding

Summary: Understanding the 3D world from images and videos is fundamental to how intelligent systems perceive and interact with their environments. However, existing approaches often struggle when visual overlap is limited or when scenes exhibit strong ambiguities, such as repetitive or symmetric structures. In this talk, I will present how learned priors can help overcome these challenges: leveraging generative video models for extreme camera pose estimation, and identifying ambiguous scene matches that hinder accurate 3D reconstruction. These approaches take a step toward more resilient and generalizable 3D understanding, and I will conclude by outlining how foundation models may open new possibilities for integrating perception, reasoning, and spatial intelligence.