Main Page

From NSL-Reading

Jump to: navigation, search

Neural Systems Lab Reading Group

Fridays 12:30 in room 303


Sessions

  • Feb 6

(tentative - open to suggestions beforehand) Reinforcement Learning Overview as presented by Douglas Aberdeen videos here


  • Feb 13

Lecture 1 http://ocw.mit.edu/OcwWeb/Mechanical-Engineering/2-997Spring2004/CourseHome/index.htm

Markov Decision Processes

Finite-Horizon Problems: Backwards Induction

Discounted-Cost Problems: Cost-to-Go Function, Bellman's Equation


  • Feb 20

Do-over! We're doing lecture 1 again.

Lecture 1 http://ocw.mit.edu/OcwWeb/Mechanical-Engineering/2-997Spring2004/CourseHome/index.htm


  • Feb 27

Value Iteration

Existence and Uniqueness of Bellman's Equation Solution

Gauss-Seidel Value Iteration

Lecture Notes Here


  • Mar 6

Lecture notes here

Topics

Feel free to add

  • Hierarchical RL
  • The basics! Specific reinforcement learning techniques: DP, TD(lambda) , Monte

Carlo methods, Kalman Filters in RL context, RTDP

  • RL and neuroscience (See Dayan and Daw 2008)
  • POMDP approximation techniques
  • Demonstrations of successful and unsuccessful applications of RL
  • Models of visual attention and search
  • Cognitive architectures and models of specific cognitive tasks
  • State representation, including why continuous domains are so deadly, and what is the right way to cope?

Wiki Information Which will eventually be removed

Retrieved from "http://neural.cs.washington.edu/NSL-Reading/Main_Page"