Iterative Monte Carlo Tree Search for Neural Architecture Search

Mehraveh Javan Roshtkhari, Matthew Toews, Marco Pedersoli
Proceedings of the Fourth International Conference on Automated Machine Learning, PMLR 293:3/1-17, 2025.

Abstract

Recent work has shown Monte-Carlo Tree Search (MCTS) as an effective approach for Neural Architecture Search (NAS) in producing competitive architectures. However, the performance of the tree search is highly sensitive to the node visiting order. If the initial nodes are highly discriminative, good configurations can be efficiently found with minimal sampling. In contrast, non-discriminative initial nodes require exploring an exponential number of nodes before finding good solutions. In this paper, we present an iterative NAS approach to jointly train the recognition model with MCTS and learn the optimal node ordering of the tree. With our approach, the order of node visits in the tree is iteratively refined based on the estimated performance of the nodes on the validation set. With this approach, good architectures are more likely to naturally emerge at the beginning of the tree, improving the search process. Experiments on two classification benchmarks and a segmentation task show that the proposed method can improve the performance of MCTS, compared to state-of-the-art MCTS approaches for NAS.

Cite this Paper

BibTeX
@InProceedings{pmlr-v293-roshtkhari25a, title = {Iterative Monte Carlo Tree Search for Neural Architecture Search}, author = {Roshtkhari, Mehraveh Javan and Toews, Matthew and Pedersoli, Marco}, booktitle = {Proceedings of the Fourth International Conference on Automated Machine Learning}, pages = {3/1--17}, year = {2025}, editor = {Akoglu, Leman and Doerr, Carola and van Rijn, Jan N. and Garnett, Roman and Gardner, Jacob R.}, volume = {293}, series = {Proceedings of Machine Learning Research}, month = {08--11 Sep}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v293/main/assets/roshtkhari25a/roshtkhari25a.pdf}, url = {https://proceedings.mlr.press/v293/roshtkhari25a.html}, abstract = {Recent work has shown Monte-Carlo Tree Search (MCTS) as an effective approach for Neural Architecture Search (NAS) in producing competitive architectures. However, the performance of the tree search is highly sensitive to the node visiting order. If the initial nodes are highly discriminative, good configurations can be efficiently found with minimal sampling. In contrast, non-discriminative initial nodes require exploring an exponential number of nodes before finding good solutions. In this paper, we present an iterative NAS approach to jointly train the recognition model with MCTS and learn the optimal node ordering of the tree. With our approach, the order of node visits in the tree is iteratively refined based on the estimated performance of the nodes on the validation set. With this approach, good architectures are more likely to naturally emerge at the beginning of the tree, improving the search process. Experiments on two classification benchmarks and a segmentation task show that the proposed method can improve the performance of MCTS, compared to state-of-the-art MCTS approaches for NAS.} }
Endnote
%0 Conference Paper %T Iterative Monte Carlo Tree Search for Neural Architecture Search %A Mehraveh Javan Roshtkhari %A Matthew Toews %A Marco Pedersoli %B Proceedings of the Fourth International Conference on Automated Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Leman Akoglu %E Carola Doerr %E Jan N. van Rijn %E Roman Garnett %E Jacob R. Gardner %F pmlr-v293-roshtkhari25a %I PMLR %P 3/1--17 %U https://proceedings.mlr.press/v293/roshtkhari25a.html %V 293 %X Recent work has shown Monte-Carlo Tree Search (MCTS) as an effective approach for Neural Architecture Search (NAS) in producing competitive architectures. However, the performance of the tree search is highly sensitive to the node visiting order. If the initial nodes are highly discriminative, good configurations can be efficiently found with minimal sampling. In contrast, non-discriminative initial nodes require exploring an exponential number of nodes before finding good solutions. In this paper, we present an iterative NAS approach to jointly train the recognition model with MCTS and learn the optimal node ordering of the tree. With our approach, the order of node visits in the tree is iteratively refined based on the estimated performance of the nodes on the validation set. With this approach, good architectures are more likely to naturally emerge at the beginning of the tree, improving the search process. Experiments on two classification benchmarks and a segmentation task show that the proposed method can improve the performance of MCTS, compared to state-of-the-art MCTS approaches for NAS.
APA
Roshtkhari, M.J., Toews, M. & Pedersoli, M.. (2025). Iterative Monte Carlo Tree Search for Neural Architecture Search. Proceedings of the Fourth International Conference on Automated Machine Learning, in Proceedings of Machine Learning Research 293:3/1-17 Available from https://proceedings.mlr.press/v293/roshtkhari25a.html.

Related Material