Research map / Probabilistic and scientific modeling / Variational and generative sampling

Bridging the Gap Between f-divergences and Bayes Hilbert Spaces

ICLR 2025 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Likelihood-free and graphical model estimation

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Generalizes f-divergences to locally non-convex generating functions, defining pseudo f-divergences, and focuses on one obtained from the induced metric of Bayes Hilbert spaces.

Problem

Traditional f-divergences require convex divergence-generating functions and cannot capture a broad range of distances between distributions.

Approach

Extends f-divergences with locally non-convex divergence-generating functions, yielding a new class of pseudo f-divergences, and studies one such pseudo divergence given by the metric induced by Bayes Hilbert spaces.

Experimental setup

Builds on: f-divergence

Closest papers by meaning

• Function-space regularized Rényi divergences (ICLR 2023 · similarity 0.90)
• Moreau-Yosida $f$-divergences (ICML 2021 · similarity 0.89)
• $f$-Divergence Based Classification: Beyond the Use of Cross-Entropy (ICML 2024 · similarity 0.87)
• Mutual Information Estimation via $f$-Divergence and Data Derangements (NeurIPS 2024 · similarity 0.86)
• Stochastic variance-reduced Gaussian variational inference on the Bures-Wasserstein manifold (ICLR 2025 · similarity 0.85)
• Bounds on $L_p$ Errors in Density Ratio Estimation via $f$-Divergence Loss Functions (ICLR 2025 · similarity 0.85)