VDebugger: Harnessing Execution Feedback for Debugging Visual Programs

TL;DR

VDebugger is a new framework that uses execution feedback to identify and fix logic errors in visual programs generated by language models, significantly improving their accuracy and generalization.

Contribution

It introduces a critic-refiner framework trained to debug visual programs by tracking execution step-by-step, a novel error injection pipeline, and demonstrates effectiveness across multiple datasets.

Findings

Up to 3.2% improvement in downstream task accuracy

Generalizes well to unseen tasks with 2.3% improvement

Effective in localizing and correcting logic errors in visual programs

Abstract

Visual programs are executable code generated by large language models to address visual reasoning problems. They decompose complex questions into multiple reasoning steps and invoke specialized models for each step to solve the problems. However, these programs are prone to logic errors, with our preliminary evaluation showing that 58% of the total errors are caused by program logic errors. Debugging complex visual programs remains a major bottleneck for visual reasoning. To address this, we introduce VDebugger, a novel critic-refiner framework trained to localize and debug visual programs by tracking execution step by step. VDebugger identifies and corrects program errors leveraging detailed execution feedback, improving interpretability and accuracy. The training data is generated through an automated pipeline that injects errors into correct visual programs using a novel mask-best decoding technique. Evaluations on six datasets demonstrate VDebugger's effectiveness, showing performance improvements of up to 3.2% in downstream task accuracy. Further studies show VDebugger's ability to generalize to unseen tasks, bringing a notable improvement of 2.3% on the unseen COVR task. Code, data and models are made publicly available at https://github.com/shirley-wu/vdebugger/