{"ID":2891978,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2507.17013","arxiv_id":"2507.17013","title":"laplax -- Laplace Approximations with JAX","abstract":"The Laplace approximation provides a scalable and efficient means of quantifying weight-space uncertainty in deep neural networks, enabling the application of Bayesian tools such as predictive uncertainty and model selection via Occam's razor. In this work, we introduce laplax, a new open-source Python package for performing Laplace approximations with jax. Designed with a modular and purely functional architecture and minimal external dependencies, laplax offers a flexible and researcher-friendly framework for rapid prototyping and experimentation. Its goal is to facilitate research on Bayesian neural networks, uncertainty quantification for deep learning, and the development of improved Laplace approximation techniques.","short_abstract":"The Laplace approximation provides a scalable and efficient means of quantifying weight-space uncertainty in deep neural networks, enabling the application of Bayesian tools such as predictive uncertainty and model selection via Occam's razor. In this work, we introduce laplax, a new open-source Python package for perf...","url_abs":"https://arxiv.org/abs/2507.17013","url_pdf":"https://arxiv.org/pdf/2507.17013v1","authors":"[\"Tobias Weber\",\"Bálint Mucsányi\",\"Lenard Rommel\",\"Thomas Christie\",\"Lars Kasüschke\",\"Marvin Pförtner\",\"Philipp Hennig\"]","published":"2025-07-22T20:49:30Z","proceeding":"cs.LG","tasks":"[\"cs.LG\",\"cs.AI\"]","methods":"[]","has_code":false}