VDC: Versatile Data Cleanser based on Visual-Linguistic Inconsistency by Multimodal Large Language Models

TL;DR

This paper introduces VDC, a novel data cleansing method using multimodal large language models to detect various dirty samples in datasets by identifying visual-linguistic inconsistencies, enhancing data quality for AI systems.

Contribution

VDC leverages multimodal large language models to detect diverse dirty samples through visual-linguistic inconsistency analysis, offering improved generalization over existing methods.

Findings

VDC outperforms existing detectors in identifying poisoned and noisy samples.

VDC generalizes well across different dirty sample categories.

Extensive experiments confirm VDC's superior detection capabilities.

Abstract

The role of data in building AI systems has recently been emphasized by the emerging concept of data-centric AI. Unfortunately, in the real-world, datasets may contain dirty samples, such as poisoned samples from backdoor attack, noisy labels in crowdsourcing, and even hybrids of them. The presence of such dirty samples makes the DNNs vunerable and unreliable.Hence, it is critical to detect dirty samples to improve the quality and realiability of dataset. Existing detectors only focus on detecting poisoned samples or noisy labels, that are often prone to weak generalization when dealing with dirty samples from other domains.In this paper, we find a commonality of various dirty samples is visual-linguistic inconsistency between images and associated labels. To capture the semantic inconsistency between modalities, we propose versatile data cleanser (VDC) leveraging the surpassing capabilities of multimodal large language models (MLLM) in cross-modal alignment and reasoning.It consists of three consecutive modules: the visual question generation module to generate insightful questions about the image; the visual question answering module to acquire the semantics of the visual content by answering the questions with MLLM; followed by the visual answer evaluation module to evaluate the inconsistency.Extensive experiments demonstrate its superior performance and generalization to various categories and types of dirty samples. The code is available at \url{https://github.com/zihao-ai/vdc}.