Conf-Gen: Conformal Uncertainty Quantification for Generative Models

Gabriel Loaiza-Ganem, Kevin Zhang, Wei Cui, Marc T. Law, Kin Kwan Leung
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:80383-80412, 2026.

Abstract

Conformal prediction (CP) and its extension, conformal risk control (CRC), are established frameworks for quantifying uncertainty in supervised machine learning through formal guarantees. However, recent breakthroughs in artificial intelligence (AI) have been driven by unsupervised generative models, such as large language models (LLMs) and image generators, which are not directly compatible with CP or CRC. In this work we introduce conformal generation (Conf-Gen), a general framework adapting CRC to generative tasks while relaxing its theoretical assumptions. Conf-Gen unifies and generalizes previous attempts to apply CP to LLMs, and extends conformal methodology to entirely new domains. We demonstrate the flexibility of Conf-Gen through some novel applications, including obtaining conformal guarantees on: image generators producing non-memorized images, conversational AI systems having asked enough clarifying questions, and the output of AI agents being correct.

Cite this Paper


BibTeX
@InProceedings{pmlr-v306-loaiza-ganem26a, title = {Conf-Gen: Conformal Uncertainty Quantification for Generative Models}, author = {Loaiza-Ganem, Gabriel and Zhang, Kevin and Cui, Wei and Law, Marc T. and Leung, Kin Kwan}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {80383--80412}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/loaiza-ganem26a/loaiza-ganem26a.pdf}, url = {https://proceedings.mlr.press/v306/loaiza-ganem26a.html}, abstract = {Conformal prediction (CP) and its extension, conformal risk control (CRC), are established frameworks for quantifying uncertainty in supervised machine learning through formal guarantees. However, recent breakthroughs in artificial intelligence (AI) have been driven by unsupervised generative models, such as large language models (LLMs) and image generators, which are not directly compatible with CP or CRC. In this work we introduce conformal generation (Conf-Gen), a general framework adapting CRC to generative tasks while relaxing its theoretical assumptions. Conf-Gen unifies and generalizes previous attempts to apply CP to LLMs, and extends conformal methodology to entirely new domains. We demonstrate the flexibility of Conf-Gen through some novel applications, including obtaining conformal guarantees on: image generators producing non-memorized images, conversational AI systems having asked enough clarifying questions, and the output of AI agents being correct.} }
Endnote
%0 Conference Paper %T Conf-Gen: Conformal Uncertainty Quantification for Generative Models %A Gabriel Loaiza-Ganem %A Kevin Zhang %A Wei Cui %A Marc T. Law %A Kin Kwan Leung %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-loaiza-ganem26a %I PMLR %P 80383--80412 %U https://proceedings.mlr.press/v306/loaiza-ganem26a.html %V 306 %X Conformal prediction (CP) and its extension, conformal risk control (CRC), are established frameworks for quantifying uncertainty in supervised machine learning through formal guarantees. However, recent breakthroughs in artificial intelligence (AI) have been driven by unsupervised generative models, such as large language models (LLMs) and image generators, which are not directly compatible with CP or CRC. In this work we introduce conformal generation (Conf-Gen), a general framework adapting CRC to generative tasks while relaxing its theoretical assumptions. Conf-Gen unifies and generalizes previous attempts to apply CP to LLMs, and extends conformal methodology to entirely new domains. We demonstrate the flexibility of Conf-Gen through some novel applications, including obtaining conformal guarantees on: image generators producing non-memorized images, conversational AI systems having asked enough clarifying questions, and the output of AI agents being correct.
APA
Loaiza-Ganem, G., Zhang, K., Cui, W., Law, M.T. & Leung, K.K.. (2026). Conf-Gen: Conformal Uncertainty Quantification for Generative Models. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:80383-80412 Available from https://proceedings.mlr.press/v306/loaiza-ganem26a.html.

Related Material