Skip to yearly menu bar Skip to main content


Poster

A Bias-Variance-Covariance Decomposition of Kernel Scores for Generative Models

Sebastian Gregor Gruber · Florian Buettner

Hall C 4-9 #1604
[ ] [ Paper PDF ]
[ Poster
Wed 24 Jul 2:30 a.m. PDT — 4 a.m. PDT

Abstract:

Generative models, like large language models, are becoming increasingly relevant in our daily lives, yet a theoretical framework to assess their generalization behavior and uncertainty does not exist. Particularly, the problem of uncertainty estimation is commonly solved in an ad-hoc and task-dependent manner. For example, natural language approaches cannot be transferred to image generation. In this paper, we introduce the first bias-variance-covariance decomposition for kernel scores. This decomposition represents a theoretical framework from which we derive a kernel-based variance and entropy for uncertainty estimation. We propose unbiased and consistent estimators for each quantity which only require generated samples but not the underlying model itself. Based on the wide applicability of kernels, we demonstrate our framework via generalization and uncertainty experiments for image, audio, and language generation. Specifically, kernel entropy for uncertainty estimation is more predictive of performance on CoQA and TriviaQA question answering datasets than existing baselines and can also be applied to closed-source models.

Chat is not available.