eprintid: 10194864 rev_number: 8 eprint_status: archive userid: 699 dir: disk0/10/19/48/64 datestamp: 2024-07-19 14:59:46 lastmod: 2024-09-13 06:10:08 status_changed: 2024-07-19 14:59:46 type: article metadata_visibility: show sword_depositor: 699 creators_name: Lopes, Vasco creators_name: Carlucci, Fabio Maria creators_name: Esperanca, Pedro M creators_name: Singh, Marco creators_name: Yang, Antoine creators_name: Gabillon, Victor creators_name: Xu, Hang creators_name: Chen, Zewei creators_name: Wang, Jun title: Manas: multi-agent neural architecture search ispublished: pub divisions: UCL divisions: B04 divisions: F48 keywords: AutoML, Computer Science, Computer Science, Artificial Intelligence, Computer vision, Multi arm bandits, Neural architecture search, Object recognition, Science & Technology, Technology note: This version is the author accepted manuscript. For information on re-use, please refer to the publisher’s terms and conditions. abstract: The Neural Architecture Search (NAS) problem is typically formulated as a graph search problem where the goal is to learn the optimal operations over edges in order to maximize a graph-level global objective. Due to the large architecture parameter space, efficiency is a key bottleneck preventing NAS from its practical use. In this work, we address the issue by framing NAS as a multi-agent problem where agents control a subset of the network and coordinate to reach optimal architectures. We provide two distinct lightweight implementations, with reduced memory requirements (1/8th of state-of-the-art), and performances above those of much more computationally expensive methods. Theoretically, we demonstrate vanishing regrets of the form O(T) , with T being the total number of rounds. Finally, we perform experiments on CIFAR-10 and ImageNet, and aware that random search and random sampling are (often ignored) effective baselines, we conducted additional experiments on 3 alternative datasets, with complexity constraints, and 2 network configurations, and achieve competitive results in comparison with the baselines and other methods. date: 2024-01 date_type: published publisher: Springer Verlag official_url: http://dx.doi.org/10.1007/s10994-023-06379-w oa_status: green full_text_type: other language: eng primo: open primo_central: open_green verified: verified_manual elements_id: 2090757 doi: 10.1007/s10994-023-06379-w lyricists_name: Wang, Jun lyricists_id: JWANG00 actors_name: Wang, Jun actors_id: JWANG00 actors_role: owner funding_acknowledgements: 2020.04588.BD [FCT - Fundacao para a Ciencia e Tecnologia]; [Huawei Technologies R amp;D (UK) Ltd] full_text_status: public publication: Machine Learning volume: 113 number: 1 pagerange: 73-96 pages: 24 issn: 0885-6125 citation: Lopes, Vasco; Carlucci, Fabio Maria; Esperanca, Pedro M; Singh, Marco; Yang, Antoine; Gabillon, Victor; Xu, Hang; ... Wang, Jun; + view all <#> Lopes, Vasco; Carlucci, Fabio Maria; Esperanca, Pedro M; Singh, Marco; Yang, Antoine; Gabillon, Victor; Xu, Hang; Chen, Zewei; Wang, Jun; - view fewer <#> (2024) Manas: multi-agent neural architecture search. Machine Learning , 113 (1) pp. 73-96. 10.1007/s10994-023-06379-w <https://doi.org/10.1007/s10994-023-06379-w>. Green open access document_url: https://discovery.ucl.ac.uk/id/eprint/10194864/1/1909.01051v4.pdf