English

Concepts or Skills? Rethinking Instruction Selection for Multi-modal Models

Computer Vision and Pattern Recognition 2025-08-15 v1 Machine Learning

Abstract

Vision-language instruction tuning achieves two main purposes: learning visual concepts and learning visual skills. In this paper, we found that vision-language benchmarks fall into the dichotomy of mainly benefiting from training on instructions with similar skills or visual concepts. Inspired by the discovery, we designed a simple targeted training data selection method to optimize the performance of a given benchmark. We first extract the concepts/skills from the benchmark, determine whether the benchmark predominantly benefits from similar concepts or skills, and finally select instructions with the most matching concepts/skills. Experiments on 10+ benchmarks validate the effectiveness of our targeted data selection method, showing +0.9\% over the best existing baseline averaged over all benchmarks and +1.5\% on the skill-focused subset. Our findings underscore the importance of recognizing the inherent trade-off within instruction selection, which requires balancing the acquisition of conceptual knowledge against visual skill.

Keywords

Cite

@article{arxiv.2508.10339,
  title  = {Concepts or Skills? Rethinking Instruction Selection for Multi-modal Models},
  author = {Andrew Bai and Justin Cui and Ruochen Wang and Cho-Jui Hsieh},
  journal= {arXiv preprint arXiv:2508.10339},
  year   = {2025}
}

Comments

11 pages, 1 figure

R2 v1 2026-07-01T04:49:16.244Z