Will It Zero-Shot?: Predicting Zero-Shot Classification Performance For Arbitrary Queries

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Robbins, Kevin, Liu, Xiaotong, Wu, Yu, Sun, Le, McPeak, Grady, Stylianou, Abby, Pless, Robert
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910071787618304
author Robbins, Kevin
Liu, Xiaotong
Wu, Yu
Sun, Le
McPeak, Grady
Stylianou, Abby
Pless, Robert
author_facet Robbins, Kevin
Liu, Xiaotong
Wu, Yu
Sun, Le
McPeak, Grady
Stylianou, Abby
Pless, Robert
contents Vision-Language Models like CLIP create aligned embedding spaces for text and images, making it possible for anyone to build a visual classifier by simply naming the classes they want to distinguish. However, a model that works well in one domain may fail in another, and non-expert users have no straightforward way to assess whether their chosen VLM will work on their problem. We build on prior work using text-only comparisons to evaluate how well a model works for a given natural language task, and explore approaches that also generate synthetic images relevant to that task to evaluate and refine the prediction of zero-shot accuracy. We show that generated imagery to the baseline text-only scores substantially improves the quality of these predictions. Additionally, it gives a user feedback on the kinds of images that were used to make the assessment. Experiments on standard CLIP benchmark datasets demonstrate that the image-based approach helps users predict, without any labeled examples, whether a VLM will be effective for their application.
format Preprint
id arxiv_https___arxiv_org_abs_2601_17535
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Will It Zero-Shot?: Predicting Zero-Shot Classification Performance For Arbitrary Queries
Robbins, Kevin
Liu, Xiaotong
Wu, Yu
Sun, Le
McPeak, Grady
Stylianou, Abby
Pless, Robert
Computer Vision and Pattern Recognition
Vision-Language Models like CLIP create aligned embedding spaces for text and images, making it possible for anyone to build a visual classifier by simply naming the classes they want to distinguish. However, a model that works well in one domain may fail in another, and non-expert users have no straightforward way to assess whether their chosen VLM will work on their problem. We build on prior work using text-only comparisons to evaluate how well a model works for a given natural language task, and explore approaches that also generate synthetic images relevant to that task to evaluate and refine the prediction of zero-shot accuracy. We show that generated imagery to the baseline text-only scores substantially improves the quality of these predictions. Additionally, it gives a user feedback on the kinds of images that were used to make the assessment. Experiments on standard CLIP benchmark datasets demonstrate that the image-based approach helps users predict, without any labeled examples, whether a VLM will be effective for their application.
title Will It Zero-Shot?: Predicting Zero-Shot Classification Performance For Arbitrary Queries
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2601.17535