Timeline for What are the advantages of RL with actor-critic methods over actor-only methods?
Current License: CC BY-SA 4.0
7 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 13, 2021 at 16:53 | comment | added | nbro | Ok, thanks! That makes this answer clearer. Another thing that I think is missing from this answer is that you seem to focus on the comparison of actor-only methods and critic-only methods, though it seems to be implied that all the info mentioned is applicable to actor-critic methods as well as critic-only methods (such as Q-learning). In any case, it may be a good idea to differentiate actor-critic methods and critic-only methods. | |
| Jan 13, 2021 at 16:14 | history | edited | Neil Slater | CC BY-SA 4.0 |
added 203 characters in body
|
| Jan 13, 2021 at 15:05 | comment | added | Neil Slater | @nbro I expanded the first paragraph with an explanation and linked a related question on cross-validated to address your second point | |
| Jan 13, 2021 at 15:03 | history | edited | Neil Slater | CC BY-SA 4.0 |
added 404 characters in body
|
| Jan 13, 2021 at 10:51 | comment | added | nbro | Maybe you should at least briefly (or link to some reference that) describe(s) why 1. actor-only methods (such as REINFORCE) require episodic problems and 2. TD-based methods may have lower bias. | |
| Jan 13, 2021 at 10:38 | history | edited | nbro | CC BY-SA 4.0 |
edits to reflect the edits made to the question
|
| Jan 12, 2021 at 22:59 | history | answered | Neil Slater | CC BY-SA 4.0 |