AI RESEARCH
HAVEN: Hierarchically Aligned Multimodal Benchmark for Unified Video Understanding
arXiv CS.CV
•
ArXi:2605.19223v1 Announce Type: new While Multimodal Large Language Models (MLLMs) exhibit strong performance on standard video tasks, their ability to faithfully summarize and reason over complex narratives remains poorly evaluated. Existing summarization benchmarks fragment supervision across isolated granularities, such as keyframes, key shots, or disjointed text summaries, failing to capture the inherently hierarchical structure of cross-modal alignment. To address this critical gap, we.