English subtitles

← 10ps-03 Function Generalization

Homework 5 2 Function Generalization.mp4

Get Embed Code
2 Languages

Showing Revision 1 created 11/28/2012 by Amara Bot.

  1. This question involves function generalization in reinforcement learning,
  2. and we're operating in a 1-dimensional environment of squares,
  3. and we're going to consider a state generalization function,
  4. that is a function that takes a state such as this and condenses it into some features
  5. to represent that state.
  6. The first function we're going to consider F has these features--
  7. f1 is the distance from the Agent represented by A to the goal represented by G,
  8. and f2--the distance from the Agent to the closest Bad guy
  9. which is represented by a B.
  10. So that's the function F, and we also want to consider the function G
  11. which has the same 2 features--f1 and f2--and adds a third feature
  12. which is the distance of the closest Bad guy to the goal.
  13. That is distance from the goal to the Bad guy--the minimum of that over
  14. all possible Bad guys,
  15. and now I want you to say which of the states below--these 3 states--
  16. have the same value as the state above--this state--under the functions F and G.
  17. And click off the ones that have the same, and then I want you to answer for me--
  18. In this world, agents and Bad guys can move one Square at a time,
  19. and the agent tries to get to the goal without encountering Bad guys,
  20. and for the agent to do that, which is a more useful generalization function
  21. to use over these states--F or G?