Cliqueformer: Model-Based Optimization with Structured Transformers
Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (27) · 22679-22687
DOI: 10.1609/aaai.v40i27.39429 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Jakub Grudzien Kuba und 2 weiteren Autor:innen.
Abstract (Originalsprache)
Large neural networks excel at prediction tasks, but their application to design problems, such as protein engineering or materials discovery, requires solving offline model-based optimization (MBO) problems. While predictive models may not directly translate to effective design, recent MBO algorithms incorporate reinforcement learning and generative modeling approaches. Meanwhile, theoretical work suggests that exploiting the target function’s structure can enhance MBO performance. We present Cliqueformer, a transformer- based architecture that learns the black-box function’s structure through functional graphical models (FGM), addressing distribution shift without relying on explicit conservative approaches. Across various domains, including chemical and genetic design tasks, Cliqueformer demonstrates superior performance compared to existing methods.
Zitieren
APA
Kuba, J. G., Abbeel, P., & Levine, S. (2026). Cliqueformer: Model-Based Optimization with Structured Transformers. Proceedings of the AAAI Conference on Artificial Intelligence, 40(27), 22679-22687. https://doi.org/10.1609/aaai.v40i27.39429
Vancouver
Kuba JG, Abbeel P, Levine S. Cliqueformer: Model-Based Optimization with Structured Transformers. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(27):22679-22687. doi:10.1609/aaai.v40i27.39429
BibTeX
@article{Kuba2026,
author = {Kuba, Jakub Grudzien and Abbeel, Pieter and Levine, Sergey},
title = {Cliqueformer: Model-Based Optimization with Structured Transformers},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {27},
pages = {22679--22687},
doi = {10.1609/aaai.v40i27.39429}
}
EUVIMED