Jesse Cummings

Jesse  Cummings
MIT EECS | Aptiv Undergraduate Research and Innovation Scholar
Advisor: Boris Katz
Department: Electrical Engineering and Computer Science
Areas of Research: Computer Graphics and Vision
Years: 2022-2023
Research Project Title:

Improved Computer Vision Benchmarks with Human Psychophysics Experiments

abstract:Exactly how the brain solves the problem of robust object recognition is a mystery that if solved would have remarkable implications for the fields of neuroscience and AI. In this project, we investigate this question by recording the brain activity of primates as they view images that in previous work was shown to be either easy or difficult for humans to recognize. We hope to find differences in the spatial and temporal patterns of brain activity which could serve as an indication of what kind of mechanisms are exploited in order to solve difficult object recognition. Identifying motifs in the brain that are essential for robust recognition could lead to new innovations in the field of AI as we implement algorithms that can approximate human visual abilities increasingly well.

I am fascinated by intelligence. What it is, how it emerges in different organisms, and how to reproduce it in silica. I want to leverage our increasingly deep understanding of the human brain to improve AI models.