notesum.ai
Published at November 7Interpreting the Learned Model in MuZero Planning
cs.AI
cs.LG
Released Date: November 7, 2024
Authors: Hung Guei1, Yan-Ru Ju1, Wei-Yu Chen2, Ti-Rong Wu1
Aff.: 1Institute of Information Science, Academia Sinica, Taipei, Taiwan; 2Institute of Information Science, Academia Sinica, Taipei, Taiwan; Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan
| \topruleHyperparameter | Board Games | Atari Games |
|---|---|---|
| \midruleIteration | 300 | |
| Training steps | 60k | |
| Batch size | 512 | |
| Unroll steps () | 5 | |
| # Blocks | 3 | 2 |
| # Simulations | 16 | 18 |
| Decoder scale () | 1 | 25 |
| Consistency scale () | 0 | 1 |
| \bottomrule | ||