CausalChaos! Dataset for Comprehensive Causal Action Question Answering Over Longer Causal Chains Grounded in Dynamic Visual Scenes

TL;DR

CausalChaos! is a new challenging dataset based on cartoons that enables research on complex causal reasoning in visual question answering, emphasizing multi-level causal chains and explanations in dynamic scenes.

Contribution

We introduce CausalChaos!, a novel dataset for causal video question answering with multi-level causal questions, grounded in animated scenes, and include hard incorrect answer mining to challenge models.

Findings

Models perform reasonably but need improvement on open-ended answers.

The dataset reveals the need for advanced causal relationship modeling.

Grounded in cartoon scenes, it offers clear causal signals for reasoning.

Abstract

Causal video question answering (QA) has garnered increasing interest, yet existing datasets often lack depth in causal reasoning. To address this gap, we capitalize on the unique properties of cartoons and construct CausalChaos!, a novel, challenging causal Why-QA dataset built upon the iconic "Tom and Jerry" cartoon series. Cartoons use the principles of animation that allow animators to create expressive, unambiguous causal relationships between events to form a coherent storyline. Utilizing these properties, along with thought-provoking questions and multi-level answers (answer and detailed causal explanation), our questions involve causal chains that interconnect multiple dynamic interactions between characters and visual scenes. These factors demand models to solve more challenging, yet well-defined causal relationships. We also introduce hard incorrect answer mining, including a causally confusing version that is even more challenging. While models perform well, there is much room for improvement, especially, on open-ended answers. We identify more advanced/explicit causal relationship modeling & joint modeling of vision and language as the immediate areas for future efforts to focus upon. Along with the other complementary datasets, our new challenging dataset will pave the way for these developments in the field.