Timeline for In which community does using a Bayesian regression model as a reward function with exploration vs. exploitation challenges fall under?
Current License: CC BY-SA 4.0
25 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 24, 2023 at 16:38 | history | edited | paul | CC BY-SA 4.0 |
**EDIT** part where i tie this problem together with another thread i created
|
| Jul 23, 2023 at 12:08 | vote | accept | paul | ||
| Jul 22, 2023 at 19:26 | history | edited | DeepQZero | CC BY-SA 4.0 |
edited grammar and enhanced clarity
|
| Jul 22, 2023 at 19:08 | answer | added | DeepQZero | timeline score: 1 | |
| Jul 22, 2023 at 17:27 | comment | added | DeepQZero | I finally got some time to look at this question again. I'll try to give an answer soon, perhaps today. | |
| Jul 12, 2023 at 15:40 | comment | added | paul | @DeepQZero hi, hope you are doing well, just wanted to send a little reminder in case you had forgotten, i appreciate your help immensely. kind regards, Paul | |
| Jul 7, 2023 at 18:22 | comment | added | paul | @DeepQZero no worries, i am just extremely happy someone can actually answer some of my questions, no matter how long it takes :), i created a separate thread/question(read the EDIT part in this question), that sums up my main concerns that popped up in my head when digesting the sOED reports, i would be so happy if you could shed some light on my thoughts in that thread(it is basically a bridge between an algorithm stemming from the problem above and sOED reports): ai.stackexchange.com/questions/41198/… | |
| Jul 7, 2023 at 17:52 | history | edited | paul | CC BY-SA 4.0 |
added 550 characters in body
|
| Jul 6, 2023 at 13:56 | comment | added | DeepQZero | Glad to hear it! I'm busy until the weekend (at the very earliest) anyways, so I can't look at any of this until at least then. This sOED problem is very rich, so it doesn't surprise me that it's prompting more questions in your mind. | |
| Jul 5, 2023 at 11:29 | comment | added | paul | @DeepQZero wow, finally i found reports tackling something similar, awesome reports, not just that one, but all the other sOED reports. Will definitely reference them in my final report about my real problem(the one presented here is just a toy one showcasing main considerations). A wide range of questions popped up in my mind as i am digesting the reports and trying to bridge it to my problem. I aim to create maybe separate threads showcasing these questions by friday approx and then link those threads in this one. Hold back until i create those threads, i think you will ace all of them. | |
| Jun 30, 2023 at 19:15 | comment | added | DeepQZero | Try this arxiv paper: Bayesian Sequential Optimal Experimental Design for Nonlinear Models Using Policy Gradient Reinforcement Learning | |
| Jun 30, 2023 at 19:09 | comment | added | paul | @DeepQZero Lovely, thank you very much. Are there any specific research papers or Google searches that you recommend right away? | |
| Jun 30, 2023 at 19:04 | comment | added | DeepQZero | I'll try to give an answer by the end of next week. If I don't give an answer or don't give an update here, please send me another message in this comment chain to get my attention again. | |
| Jun 29, 2023 at 17:59 | comment | added | paul | i would say that certain bandit-problems, active learning for regression, MPC is the closest i got to finding research that resembles this problem but they all have parts that are very different from this setting | |
| Jun 29, 2023 at 16:01 | comment | added | paul | @DeepQZero it is solely helpful in maximizing the total return. No extra reward added for gaining info about them. | |
| Jun 29, 2023 at 14:14 | comment | added | DeepQZero | Thank you for the update. Another question - is gaining info about the model parameters rewarded somehow, or is it only helpful in maximizing the total return? I'm asking this question since I usually have seen maximizing total rewards as the sole goal in similar problems, whereas gaining info about the model parameters is only a byproduct of trying to maximize the total rewards. If you need clarification, just let me know. | |
| Jun 29, 2023 at 8:28 | history | edited | paul | CC BY-SA 4.0 |
added 527 characters in body
|
| Jun 29, 2023 at 8:24 | comment | added | paul | also added a couple of more strategies(policies) i been thinking off, it seems so strange however that this problem has not been tackled before extensively in the literature. | |
| Jun 29, 2023 at 8:22 | history | edited | paul | CC BY-SA 4.0 |
added 527 characters in body
|
| Jun 29, 2023 at 7:16 | comment | added | paul | i would love to discuss this problem with someone who actually worked on something similar @DeepQZero because everywhere i look it is extremely difficult to find relevant info on this, bits and pieces can be taken from all the sequential decision communities but it seems so strange that noone had tackled this exact problem type before | |
| Jun 29, 2023 at 7:14 | comment | added | paul | hi @DeepQZero , just edited: My main question is, where can i find more information about my problem type, under which communities, has this problem-type even been tackled before? | |
| Jun 29, 2023 at 7:13 | history | edited | paul | CC BY-SA 4.0 |
added 106 characters in body
|
| Jun 28, 2023 at 17:04 | comment | added | DeepQZero | Hi @paul and welcome to AI Stack Exchange! If possible, please outline your specific question in the body of the post. Right now, I'm unsure of what question to answer. I did my PhD in some very similar topic, and I'd like to help out if possible. Thanks, and we look forward to more of your questions in the future! | |
| S Jun 26, 2023 at 8:14 | review | First questions | |||
| Jun 27, 2023 at 16:47 | |||||
| S Jun 26, 2023 at 8:14 | history | asked | paul | CC BY-SA 4.0 |