Memory-enhanced Retrieval Augmentation for Long Video Understanding
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866909653504360448 |
|---|---|
| author | Yuan, Huaying Liu, Zheng Qin, Minghao Qian, Hongjin Shu, Yan Dou, Zhicheng Wen, Ji-Rong Sebe, Nicu |
| author_facet | Yuan, Huaying Liu, Zheng Qin, Minghao Qian, Hongjin Shu, Yan Dou, Zhicheng Wen, Ji-Rong Sebe, Nicu |
| contents | Efficient long-video understanding~(LVU) remains a challenging task in computer vision. Current long-context vision-language models~(LVLMs) suffer from information loss due to compression and brute-force downsampling. While retrieval-augmented generation (RAG) methods mitigate this issue, their applicability is limited due to explicit query dependency. To overcome this challenge, we introduce a novel memory-enhanced RAG-based approach called MemVid, which is inspired by the cognitive memory of human beings. Our approach operates in four basic steps: 1) memorizing holistic video information, 2) reasoning about the task's information needs based on memory, 3) retrieving critical moments based on the information needs, and 4) focusing on the retrieved moments to produce the final answer. To enhance the system's memory-grounded reasoning capabilities while achieving optimal end-to-end performance, we propose a curriculum learning strategy. This approach begins with supervised learning on well-annotated reasoning results, then progressively explores and reinforces more plausible reasoning outcomes through reinforcement learning. We perform extensive evaluations on popular LVU benchmarks, including MLVU, VideoMME and LVBench. In our experiments, MemVid demonstrates superior efficiency and effectiveness compared to both LVLMs and RAG methods. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_09149 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Memory-enhanced Retrieval Augmentation for Long Video Understanding Yuan, Huaying Liu, Zheng Qin, Minghao Qian, Hongjin Shu, Yan Dou, Zhicheng Wen, Ji-Rong Sebe, Nicu Computer Vision and Pattern Recognition Multimedia Efficient long-video understanding~(LVU) remains a challenging task in computer vision. Current long-context vision-language models~(LVLMs) suffer from information loss due to compression and brute-force downsampling. While retrieval-augmented generation (RAG) methods mitigate this issue, their applicability is limited due to explicit query dependency. To overcome this challenge, we introduce a novel memory-enhanced RAG-based approach called MemVid, which is inspired by the cognitive memory of human beings. Our approach operates in four basic steps: 1) memorizing holistic video information, 2) reasoning about the task's information needs based on memory, 3) retrieving critical moments based on the information needs, and 4) focusing on the retrieved moments to produce the final answer. To enhance the system's memory-grounded reasoning capabilities while achieving optimal end-to-end performance, we propose a curriculum learning strategy. This approach begins with supervised learning on well-annotated reasoning results, then progressively explores and reinforces more plausible reasoning outcomes through reinforcement learning. We perform extensive evaluations on popular LVU benchmarks, including MLVU, VideoMME and LVBench. In our experiments, MemVid demonstrates superior efficiency and effectiveness compared to both LVLMs and RAG methods. |
| title | Memory-enhanced Retrieval Augmentation for Long Video Understanding |
| topic | Computer Vision and Pattern Recognition Multimedia |
| url | https://arxiv.org/abs/2503.09149 |