An Introduction to Reinforcement Learning and RL Brain Machine Interface

This month we have Aditya Tarigoppula from SUNY medical center presenting "An Introduction to Reinforcement Learning and RL Brain Machine Interface". Here's his abstract and bio:

Abstract:

Reinforcement learning (RL) treads a path between supervised and unsupervised learning. Similar to the daily decision process undertaken by us to perform actions, the "agent" under the RL architecture is provided with a task to decide on actions in an "environment" to maximize the cumulative "reward". Since our birth we utilize RL to shape our image of the environment, to assign values (reward) to these variables in the environment which heavily guides our action selection in a given state. RL can be used to drive a car automatically, to fly a helicphter and if seen broadly the theory allows an user to learn and perform any given task. Following an introduction to the field of RL, I will review our intented application of different models of RL to the field of Brain Machine Interface (BMI).

Bio:

I have been working at SUNY medical ceter in Dr. Joseph Francis lab (http://www.joefrancislab.com (http://www.joefrancislab.com/)) for the past one and half years wherein we have been successful in developing a reinforcement learning based brain machine interface. My current focus is on applying different models of the RL-BMI and understanding the co-adaptation that occurs between a learning user and an adaptive decoder. I did my Bachelors and Masters in Biomedical Engineering from Gujarat University, India and NYU-Poly respectively.