Memory-Augmented Monte Carlo Tree Search

Authors: Chenjun Xiao, Jincheng Mei, Martin Müller

AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate MMCTS in the game of Go. Experimental results show that MMCTS outperforms the original MCTS with the same number of simulations.
Researcher Affiliation Academia Chenjun Xiao, Jincheng Mei, Martin Müller Computing Science, University of Alberta Edmonton, Canada {chenjun,jmei2,mmueller}@ualberta.ca
Pseudocode No The paper does not contain structured pseudocode or algorithm blocks (clearly labeled algorithm sections or code-like formatted procedures).
Open Source Code No The paper does not provide concrete access to source code (specific repository link, explicit code release statement, or code in supplementary materials) for the methodology described in this paper.
Open Datasets Yes Our implementation applies a deep convolutional neural network (DCNN) from (Clark and Storkey 2015), which is trained for move prediction by professional game records.
Dataset Splits No The paper does not provide specific dataset split information (exact percentages, sample counts, citations to predefined splits, or detailed splitting methodology) needed to reproduce the data partitioning.
Hardware Specification No The paper does not provide specific hardware details (exact GPU/CPU models, processor types with speeds, memory amounts, or detailed computer specifications) used for running its experiments.
Software Dependencies No The paper does not provide specific ancillary software details (e.g., library or solver names with version numbers like Python 3.8, CPLEX 12.4) needed to replicate the experiment.
Experiment Setup Yes The feature hashing dimension is set to 4096... The hash code in our Sim Hash implementation has 16 bits. We use 8 hash tables... set to 1 in our implementation. The discount parameter η in equation (9)... is set to 2... We set CNN STRENGTH to 200 in our experiment... we restrict DCNN calls to the first 100 nodes visited during the search... The parameter M is chosen from {20, 50, 100}, and τ from {0.05, 0.1, 1}. The size of M is set to 10000. We vary the number of simulations per move from {1000, 5000, 10000}.