MECD: Unlocking Multi-Event Causal Discovery in Video Reasoning

TL;DR

This paper introduces MECD, a new task and dataset for discovering causal relationships among multiple events in long videos, employing a novel framework inspired by Granger Causality and causal inference techniques.

Contribution

The paper presents a new multi-event causal discovery task, dataset, and a novel framework combining event prediction with causal inference methods for video reasoning.

Findings

Outperforms GPT-4o and VideoLLaVA in causal discovery accuracy.

Effectively uncovers causal relationships in long, multi-event videos.

Demonstrates the utility of combining Granger Causality with causal inference techniques.

Abstract

Video causal reasoning aims to achieve a high-level understanding of video content from a causal perspective. However, current video reasoning tasks are limited in scope, primarily executed in a question-answering paradigm and focusing on short videos containing only a single event and simple causal relationships, lacking comprehensive and structured causality analysis for videos with multiple events. To fill this gap, we introduce a new task and dataset, Multi-Event Causal Discovery (MECD). It aims to uncover the causal relationships between events distributed chronologically across long videos. Given visual segments and textual descriptions of events, MECD requires identifying the causal associations between these events to derive a comprehensive, structured event-level video causal diagram explaining why and how the final result event occurred. To address MECD, we devise a novel framework inspired by the Granger Causality method, using an efficient mask-based event prediction model to perform an Event Granger Test, which estimates causality by comparing the predicted result event when premise events are masked versus unmasked. Furthermore, we integrate causal inference techniques such as front-door adjustment and counterfactual inference to address challenges in MECD like causality confounding and illusory causality. Experiments validate the effectiveness of our framework in providing causal relationships in multi-event videos, outperforming GPT-4o and VideoLLaVA by 5.7% and 4.1%, respectively.