RTime-QA: A Benchmark for Atomic Temporal Event Understanding in Large Multi-modal Models
This addresses the need for better benchmarks to assess temporal understanding in video-language AI models, though it is incremental as it builds on existing evaluation frameworks.
The paper tackles the problem of evaluating Large Multi-modal Models' (LMMs) atomic temporal event understanding by introducing RTime-QA, a benchmark with 822 video-text questions, where the state-of-the-art model Qwen2-VL initially scores only 34.6 on strict-ACC, and they show that fine-tuning with their RTime-IT dataset improves this to 65.9.
Understanding accurate atomic temporal event is essential for video comprehension. However, current video-language benchmarks often fall short to evaluate Large Multi-modal Models' (LMMs) temporal event understanding capabilities, as they can be effectively addressed using image-language models. In this paper, we introduce RTime-QA, a novel benchmark specifically designed to assess the atomic temporal event understanding ability of LMMs. RTime-QA comprises 822 high-quality, carefully-curated video-text questions, each meticulously annotated by human experts. Each question features a video depicting an atomic temporal event, paired with both correct answers and temporal negative descriptions, specifically designed to evaluate temporal understanding. To advance LMMs' temporal event understanding ability, we further introduce RTime-IT, a 14k instruction-tuning dataset that employs a similar annotation process as RTime-QA. Extensive experimental analysis demonstrates that RTime-QA presents a significant challenge for LMMs: the state-of-the-art model Qwen2-VL achieves only 34.6 on strict-ACC metric, substantially lagging behind human performance. Furthermore, our experiments reveal that RTime-IT effectively enhance LMMs' capacity in temporal understanding. By fine-tuning on RTime-IT, our Qwen2-VL achieves 65.9 on RTime-QA.