Hierarchical Long Video Understanding with Audiovisual Entity Cohesion and Agentic Search

2026-01-20

AI Summary

This work addresses the challenges of information fragmentation and semantic incoherence in long-form video understanding caused by excessively long contextual spans. To this end, the authors propose HAVEN, a unified hierarchical framework that integrates audio-visual entity consistency modeling with an agent-driven, multi-granularity retrieval mechanism. HAVEN employs a four-level structured index—spanning global summaries, scenes, clips, and entities—combined with audio-visual multimodal fusion and entity-level representation alignment to enable dynamic cross-hierarchical retrieval and reasoning. This design effectively preserves temporal coherence and entity consistency throughout the video. Evaluated on LVBench, the method achieves 84.1% overall accuracy and 80.1% on reasoning tasks, significantly outperforming existing approaches.

Technology Category

Computer Vision: Video Understanding & Activity AnalysisKnowledge Representation and Reasoning: Geometric, Spatial, and Temporal ReasoningNatural Language Processing: Sentence-level Semantics, Textual Inference, etc.

Application Category

Search and Retrieval-Augmented AI: Web query analysis, representation and understandingSemantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMsGraph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Long video understanding presents significant challenges for vision-language models due to extremely long context windows. Existing solutions relying on naive chunking strategies with retrieval-augmented generation, typically suffer from information fragmentation and a loss of global coherence. We present HAVEN, a unified framework for long-video understanding that enables coherent and comprehensive reasoning by integrating audiovisual entity cohesion and hierarchical video indexing with agentic search. First, we preserve semantic consistency by integrating entity-level representations across visual and auditory streams, while organizing content into a structured hierarchy spanning global summary, scene, segment, and entity levels. Then we employ an agentic search mechanism to enable dynamic retrieval and reasoning across these layers, facilitating coherent narrative reconstruction and fine-grained entity tracking. Extensive experiments demonstrate that our method achieves good temporal coherence, entity consistency, and retrieval efficiency, establishing a new state-of-the-art with an overall accuracy of 84.1% on LVBench. Notably, it achieves outstanding performance in the challenging reasoning category, reaching 80.1%. These results highlight the effectiveness of structured, multimodal reasoning for comprehensive and context-consistent understanding of long-form videos.

Problem

• long video understanding
• information fragmentation
• global coherence
• vision-language models
• context windows

Innovation

• audiovisual entity cohesion
• hierarchical video indexing
• agentic search
• long video understanding
• multimodal reasoning