Iterative Self-Improvement of Vision Language Models for Image Scoring and Self-Explanation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Tanji, Naoto, Yamasaki, Toshihiko
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916775401095168
author Tanji, Naoto
Yamasaki, Toshihiko
author_facet Tanji, Naoto
Yamasaki, Toshihiko
contents Image scoring is a crucial task in numerous real-world applications. To trust a model's judgment, understanding its rationale is essential. This paper proposes a novel training method for Vision Language Models (VLMs) to generate not only image scores but also corresponding justifications in natural language. Leveraging only an image scoring dataset and an instruction-tuned VLM, our method enables self-training, utilizing the VLM's generated text without relying on external data or models. In addition, we introduce a simple method for creating a dataset designed to improve alignment between predicted scores and their textual justifications. By iteratively training the model with Direct Preference Optimization on two distinct datasets and merging them, we can improve both scoring accuracy and the coherence of generated explanations.
format Preprint
id arxiv_https___arxiv_org_abs_2506_02708
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Iterative Self-Improvement of Vision Language Models for Image Scoring and Self-Explanation
Tanji, Naoto
Yamasaki, Toshihiko
Computer Vision and Pattern Recognition
Computation and Language
Image scoring is a crucial task in numerous real-world applications. To trust a model's judgment, understanding its rationale is essential. This paper proposes a novel training method for Vision Language Models (VLMs) to generate not only image scores but also corresponding justifications in natural language. Leveraging only an image scoring dataset and an instruction-tuned VLM, our method enables self-training, utilizing the VLM's generated text without relying on external data or models. In addition, we introduce a simple method for creating a dataset designed to improve alignment between predicted scores and their textual justifications. By iteratively training the model with Direct Preference Optimization on two distinct datasets and merging them, we can improve both scoring accuracy and the coherence of generated explanations.
title Iterative Self-Improvement of Vision Language Models for Image Scoring and Self-Explanation
topic Computer Vision and Pattern Recognition
Computation and Language
url https://arxiv.org/abs/2506.02708