Enhancing Visual Programming for Visual Reasoning via Probabilistic Graphs

TL;DR

This paper introduces EVPG, a novel method that transforms the non-differentiable visual programming process into a differentiable probabilistic inference, enabling end-to-end training and significantly improving visual reasoning performance.

Contribution

The paper proposes a probabilistic graph-based approach to optimize visual programming frameworks for visual reasoning, allowing gradient-based learning using only final task labels.

Findings

Significant performance improvements on GQA, NLVRv2, and Open Images datasets.

Effective transformation of VP into a differentiable process via probabilistic graphs.

Enhanced end-to-end training capability for VP in complex VR tasks.

Abstract

Recently, Visual Programming (VP) based on large language models (LLMs) has rapidly developed and demonstrated significant potential in complex Visual Reasoning (VR) tasks. Previous works to enhance VP have primarily focused on improving the quality of LLM-generated visual programs. However, they have neglected to optimize the VP-invoked pre-trained models, which serve as modules for the visual sub-tasks decomposed from the targeted tasks by VP. The difficulty is that there are only final labels of targeted VR tasks rather than labels of sub-tasks. Besides, the non-differentiable nature of VP impedes the direct use of efficient gradient-based optimization methods to leverage final labels for end-to-end learning of the entire VP framework. To overcome these issues, we propose EVPG, a method to Enhance Visual Programming for visual reasoning via Probabilistic Graphs. Specifically, we creatively build a directed probabilistic graph according to the variable dependency relationships during the VP executing process, which reconstructs the non-differentiable VP executing process into a differentiable exact probability inference process on this directed probabilistic graph. As a result, this enables the VP framework to utilize the final labels for efficient, gradient-based optimization in end-to-end supervised learning on targeted VR tasks. Extensive and comprehensive experiments demonstrate the effectiveness and advantages of our EVPG, showing significant performance improvements for VP on three classical complex VR tasks: GQA, NLVRv2, and Open Images.