CELLO: Causal Evaluation of Large Vision-Language Models
This work addresses the lack of formal causal reasoning evaluation for LVLMs, which is crucial for applications like embodied agents, though it is incremental in providing a new dataset and prompting method.
The authors tackled the problem of evaluating causal reasoning in large vision-language models by introducing a fine-grained definition of causality and constructing the CELLO dataset with 14,094 questions across four causal levels, revealing that current models struggle but improve with a proposed prompting strategy.
Causal reasoning is fundamental to human intelligence and crucial for effective decision-making in real-world environments. Despite recent advancements in large vision-language models (LVLMs), their ability to comprehend causality remains unclear. Previous work typically focuses on commonsense causality between events and/or actions, which is insufficient for applications like embodied agents and lacks the explicitly defined causal graphs required for formal causal reasoning. To overcome these limitations, we introduce a fine-grained and unified definition of causality involving interactions between humans and/or objects. Building on the definition, we construct a novel dataset, CELLO, consisting of 14,094 causal questions across all four levels of causality: discovery, association, intervention, and counterfactual. This dataset surpasses traditional commonsense causality by including explicit causal graphs that detail the interactions between humans and objects. Extensive experiments on CELLO reveal that current LVLMs still struggle with causal reasoning tasks, but they can benefit significantly from our proposed CELLO-CoT, a causally inspired chain-of-thought prompting strategy. Both quantitative and qualitative analyses from this study provide valuable insights for future research. Our project page is at https://github.com/OpenCausaLab/CELLO.