At CVPR this year, there was another plenary speaker Jack Gallant who talked about Reverse Engineering the Human Visual System. The video is here.
Let us note the use of Blender to produce truth scenes which is really what other reserchers ( see A Probabilistic Theory of Deep Learning by Ankit B. Patel, Tan Nguyen, Richard G. Baraniuk ) are suggesting to use in order to remove nuisance parameters and improve learning abilities.
We mentioned Jack Gallant and his group's work here before on Nuit Blanche, see:
- Video: Paris Machine Learning Meetup #3
- Dreaming reconstructions
- Why am I thinking there is a need for a better dictionary learning or calibration algorithm ?
Abstract of the talk:
The human brain is the most sophisticated image processing system known, capable of impressive feats of recognition and discrimination under challenging natural conditions. Reverse-engineering the brain might enable us to design artificial systems with the same capabilities. My laboratory uses a data-driven system identification approach to tackle this reverse-engineering problem. Our approach consists of four broad stages. First, we use functional MRI to measure brain activity while people watch naturalistic movies. We divide these data into two parts, one use to fit models and one for testing model predictions. Second, we use a system identification framework (based on multiple linearizing feature spaces) to model activity measured at each point in the brain. Third, we inspect the most accurate models to understand how the brain represents low-, mid- and high-level information in the movies. Finally, we use the estimated models to decode brain activity, reconstructing the structural and semantic content in the movies. Any effort to reverse-engineer the brain is inevitably limited by the spatial and temporal resolution of brain measurements, and at this time the resolution of human brain measurements is relatively poor. Still, as measurement technology progresses this framework could inform development of biologically-inspired computer vision systems, and it could aid in development of practical new brain reading technologies.
Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.
Post a Comment