Timeline for How can policy parameterization be simpler than action-value parameterization in function approximation?
Current License: CC BY-SA 4.0
10 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 18, 2018 at 7:29 | vote | accept | Ali Ghghgh | ||
| Dec 18, 2018 at 7:21 | comment | added | shimao | @AliGh The reward is maximized at 0 because $f$ is monotonically decreasing. | |
| Dec 18, 2018 at 6:27 | comment | added | Ali Ghghgh | Nothing related to the question. Just about your last claim. Thanks | |
| Dec 18, 2018 at 4:40 | comment | added | shimao | not sure what point you're trying to make | |
| Dec 18, 2018 at 3:55 | comment | added | Ali Ghghgh | Actually, it is true when we have the reward function. Otherwise, we need to know the object. | |
| Dec 18, 2018 at 3:28 | comment | added | shimao | @AliGh the MDP formulation does not require a goal -- usually the problem is considered solved when the reward is maximized. In this case, reward is maximized when you reach $0$, since $f$ is maximized there. | |
| Dec 18, 2018 at 3:25 | comment | added | Ali Ghghgh | Ok, thanks. But how about moving down by 1 in the game. What was the goal of the game | |
| Dec 18, 2018 at 3:23 | comment | added | shimao | If you're unfamiliar with the traveling salesman problem, then that probably wasn't a helpful example. Just imagine that $f$ is a very hard function to model -- the world is filled with such functions. | |
| Dec 18, 2018 at 3:21 | comment | added | Ali Ghghgh | Can you explain more everything after "-- for example", please? | |
| Dec 18, 2018 at 2:45 | history | answered | shimao | CC BY-SA 4.0 |