CLAIR: Evaluating Image Captions with Large Language Models
Submission Type: Regular Short Paper
Submission Track: Resources and Evaluation
Submission Track 2: Theme Track: Large Language Models and the Future of NLP
Keywords: Image Captioning, Evaluation, Metrics, Large Language Models
TL;DR: Asking large language models to score generated captions against reference scores correlates strongly with human judgement.
Abstract: The evaluation of machine-generated image captions poses an interesting yet persistent challenge. Effective evaluation measures must consider numerous dimensions of similarity, including semantic relevance, visual structure, object interactions, caption diversity, and specificity. Existing highly-engineered measures attempt to capture specific aspects, but fall short in providing a holistic score that aligns closely with human judgments. Here, we propose CLAIR, a novel method that leverages the zero-shot language modeling capabilities of large language models (LLMs) to evaluate candidate captions. In our evaluations, CLAIR demonstrates a stronger correlation with human judgments of caption quality compared to existing measures. Notably, on Flickr8K-Expert, CLAIR achieves relative correlation improvements over SPICE of 39.6% and over image-augmented methods such as RefCLIP-S of 18.3%. Moreover, CLAIR provides noisily interpretable results by allowing the language model to identify the underlying reasoning behind its assigned score.
Submission Number: 3946
Loading