Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2511.12027 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911266734342144 |
|---|---|
| author | Yeo, Jeong Hun Chung, Sangyun Park, Sungjune Kim, Dae Hoe Moon, Jinyoung Ro, Yong Man |
| author_facet | Yeo, Jeong Hun Chung, Sangyun Park, Sungjune Kim, Dae Hoe Moon, Jinyoung Ro, Yong Man |
| contents | Long-video understanding remains a significant challenge for Multimodal Large Language Models (MLLMs) due to inherent token limitations and the complexity of capturing long-term temporal dependencies. Existing methods often fail to capture the global context and complex event relationships necessary for deep video reasoning. To address this, we introduce GCAgent, a novel Global-Context-Aware Agent framework that achieves comprehensive long-video understanding. Our core innovation is the Schematic and Narrative Episodic Memory. This memory structurally models events and their causal and temporal relations into a concise, organized context, fundamentally resolving the long-term dependency problem. Operating in a multi-stage Perception-Action-Reflection cycle, our GCAgent utilizes a Memory Manager to retrieve relevant episodic context for robust, context-aware inference. Extensive experiments confirm that GCAgent significantly enhances long-video understanding, achieving up to 23.5\% accuracy improvement on the Video-MME Long split over a strong MLLM baseline. Furthermore, our framework establishes state-of-the-art performance among comparable 7B-scale MLLMs, achieving 73.4\% accuracy on the Long split and the highest overall average (71.9\%) on the Video-MME benchmark, validating our agent-based reasoning paradigm and structured memory for cognitively-inspired long-video understanding. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_12027 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | GCAgent: Long-Video Understanding via Schematic and Narrative Episodic Memory Yeo, Jeong Hun Chung, Sangyun Park, Sungjune Kim, Dae Hoe Moon, Jinyoung Ro, Yong Man Computer Vision and Pattern Recognition Artificial Intelligence Long-video understanding remains a significant challenge for Multimodal Large Language Models (MLLMs) due to inherent token limitations and the complexity of capturing long-term temporal dependencies. Existing methods often fail to capture the global context and complex event relationships necessary for deep video reasoning. To address this, we introduce GCAgent, a novel Global-Context-Aware Agent framework that achieves comprehensive long-video understanding. Our core innovation is the Schematic and Narrative Episodic Memory. This memory structurally models events and their causal and temporal relations into a concise, organized context, fundamentally resolving the long-term dependency problem. Operating in a multi-stage Perception-Action-Reflection cycle, our GCAgent utilizes a Memory Manager to retrieve relevant episodic context for robust, context-aware inference. Extensive experiments confirm that GCAgent significantly enhances long-video understanding, achieving up to 23.5\% accuracy improvement on the Video-MME Long split over a strong MLLM baseline. Furthermore, our framework establishes state-of-the-art performance among comparable 7B-scale MLLMs, achieving 73.4\% accuracy on the Long split and the highest overall average (71.9\%) on the Video-MME benchmark, validating our agent-based reasoning paradigm and structured memory for cognitively-inspired long-video understanding. |
| title | GCAgent: Long-Video Understanding via Schematic and Narrative Episodic Memory |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2511.12027 |