Disentangling Mean Embeddings for Better Diagnostics of Image Generators

Sebastian Gregor Gruber, Pascal Tobias Ziegler, Florian Buettner

Keywords: Generative Models, Image Generation, Mean Embeddings, Kernel Methods, Interpretability, Explainability

TL;DR: This work shows how central kernel alignment disentangles the cosine similarity of mean embeddings for novel insights into image generator performance.

Abstract:

The evaluation of image generators remains a challenge due to the limitations of traditional metrics in providing nuanced insights into specific image regions.
This is a critical problem as not all regions of an image may be learned with similar ease.
In this work, we propose a novel approach to disentangle the cosine similarity of mean embeddings into the product of cosine similarities for individual pixel clusters via central kernel alignment.
Consequently, we can quantify the contribution of the cluster-wise performance to the overall image generation performance.
We demonstrate how this enhances the explainability and the likelihood of identifying pixel regions of model misbehavior across various real-world use cases.

Track: Main track

Submitted Paper: No

Published Paper: No

Submission Number: 44