VITATECS: A Diagnostic Dataset for Temporal Concept Understanding of Video-Language Models

TL;DR

VITATECS is a diagnostic dataset designed to evaluate and improve the temporal concept understanding of video-language models, addressing limitations of existing benchmarks that rely on static visual cues.

Contribution

The paper introduces a new dataset with a taxonomy of temporal concepts and counterfactual descriptions, enabling precise assessment of temporal understanding in VidLMs.

Findings

Current models show deficiencies in temporal understanding.

The dataset reveals the need for models to better grasp temporal aspects.

Counterfactual descriptions help disentangle static and temporal information.

Abstract

The ability to perceive how objects change over time is a crucial ingredient in human intelligence. However, current benchmarks cannot faithfully reflect the temporal understanding abilities of video-language models (VidLMs) due to the existence of static visual shortcuts. To remedy this issue, we present VITATECS, a diagnostic VIdeo-Text dAtaset for the evaluation of TEmporal Concept underStanding. Specifically, we first introduce a fine-grained taxonomy of temporal concepts in natural language in order to diagnose the capability of VidLMs to comprehend different temporal aspects. Furthermore, to disentangle the correlation between static and temporal information, we generate counterfactual video descriptions that differ from the original one only in the specified temporal aspect. We employ a semi-automatic data collection framework using large language models and human-in-the-loop annotation to obtain high-quality counterfactual descriptions efficiently. Evaluation of representative video-language understanding models confirms their deficiency in temporal understanding, revealing the need for greater emphasis on the temporal elements in video-language research.