Cross-Modal Causal Relational Reasoning for Event-Level Visual Question Answering

TL;DR

This paper introduces a novel framework called CMCIR that enhances event-level visual question answering by modeling causal relationships across visual and linguistic modalities, addressing spurious correlations and capturing event dynamics.

Contribution

The work proposes a cross-modal causal relational reasoning framework with causal intervention operations, improving the understanding of event temporality and causality in VQA tasks.

Findings

Outperforms existing methods on four event-level datasets.

Effectively discovers visual-linguistic causal structures.

Enhances robustness in event-level visual question answering.

Abstract

Existing visual question answering methods often suffer from cross-modal spurious correlations and oversimplified event-level reasoning processes that fail to capture event temporality, causality, and dynamics spanning over the video. In this work, to address the task of event-level visual question answering, we propose a framework for cross-modal causal relational reasoning. In particular, a set of causal intervention operations is introduced to discover the underlying causal structures across visual and linguistic modalities. Our framework, named Cross-Modal Causal RelatIonal Reasoning (CMCIR), involves three modules: i) Causality-aware Visual-Linguistic Reasoning (CVLR) module for collaboratively disentangling the visual and linguistic spurious correlations via front-door and back-door causal interventions; ii) Spatial-Temporal Transformer (STT) module for capturing the fine-grained interactions between visual and linguistic semantics; iii) Visual-Linguistic Feature Fusion (VLFF) module for learning the global semantic-aware visual-linguistic representations adaptively. Extensive experiments on four event-level datasets demonstrate the superiority of our CMCIR in discovering visual-linguistic causal structures and achieving robust event-level visual question answering. The datasets, code, and models are available at https://github.com/HCPLab-SYSU/CMCIR.