Dataset Distillation via Vision-Language Category Prototype

TL;DR

This paper introduces a novel dataset distillation method that incorporates vision-language prototypes, using descriptive text generated by large language models to improve the semantic coherence and generalization of distilled datasets.

Contribution

The study proposes integrating text prototypes derived from large language models into dataset distillation, enhancing semantic understanding and broadening applicability beyond image-only approaches.

Findings

Achieves state-of-the-art validation performance

Generates logically coherent images with target objects

Demonstrates robust generalization across datasets

Abstract

Dataset distillation (DD) condenses large datasets into compact yet informative substitutes, preserving performance comparable to the original dataset while reducing storage, transmission costs, and computational consumption. However, previous DD methods mainly focus on distilling information from images, often overlooking the semantic information inherent in the data. The disregard for context hinders the model's generalization ability, particularly in tasks involving complex datasets, which may result in illogical outputs or the omission of critical objects. In this study, we integrate vision-language methods into DD by introducing text prototypes to distill language information and collaboratively synthesize data with image prototypes, thereby enhancing dataset distillation performance. Notably, the text prototypes utilized in this study are derived from descriptive text information generated by an open-source large language model. This framework demonstrates broad applicability across datasets without pre-existing text descriptions, expanding the potential of dataset distillation beyond traditional image-based approaches. Compared to other methods, the proposed approach generates logically coherent images containing target objects, achieving state-of-the-art validation performance and demonstrating robust generalization. Source code and generated data are available in https://github.com/zou-yawen/Dataset-Distillation-via-Vision-Language-Category-Prototype/