BIMBA: Selective-Scan Compression for Long-Range Video Question Answering
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915196013903872 |
|---|---|
| author | Islam, Md Mohaiminul Nagarajan, Tushar Wang, Huiyu Bertasius, Gedas Torresani, Lorenzo |
| author_facet | Islam, Md Mohaiminul Nagarajan, Tushar Wang, Huiyu Bertasius, Gedas Torresani, Lorenzo |
| contents | Video Question Answering (VQA) in long videos poses the key challenge of extracting relevant information and modeling long-range dependencies from many redundant frames. The self-attention mechanism provides a general solution for sequence modeling, but it has a prohibitive cost when applied to a massive number of spatiotemporal tokens in long videos. Most prior methods rely on compression strategies to lower the computational cost, such as reducing the input length via sparse frame sampling or compressing the output sequence passed to the large language model (LLM) via space-time pooling. However, these naive approaches over-represent redundant information and often miss salient events or fast-occurring space-time patterns. In this work, we introduce BIMBA, an efficient state-space model to handle long-form videos. Our model leverages the selective scan algorithm to learn to effectively select critical information from high-dimensional video and transform it into a reduced token sequence for efficient LLM processing. Extensive experiments demonstrate that BIMBA achieves state-of-the-art accuracy on multiple long-form VQA benchmarks, including PerceptionTest, NExT-QA, EgoSchema, VNBench, LongVideoBench, and Video-MME. Code, and models are publicly available at https://sites.google.com/view/bimba-mllm. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_09590 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | BIMBA: Selective-Scan Compression for Long-Range Video Question Answering Islam, Md Mohaiminul Nagarajan, Tushar Wang, Huiyu Bertasius, Gedas Torresani, Lorenzo Computer Vision and Pattern Recognition Video Question Answering (VQA) in long videos poses the key challenge of extracting relevant information and modeling long-range dependencies from many redundant frames. The self-attention mechanism provides a general solution for sequence modeling, but it has a prohibitive cost when applied to a massive number of spatiotemporal tokens in long videos. Most prior methods rely on compression strategies to lower the computational cost, such as reducing the input length via sparse frame sampling or compressing the output sequence passed to the large language model (LLM) via space-time pooling. However, these naive approaches over-represent redundant information and often miss salient events or fast-occurring space-time patterns. In this work, we introduce BIMBA, an efficient state-space model to handle long-form videos. Our model leverages the selective scan algorithm to learn to effectively select critical information from high-dimensional video and transform it into a reduced token sequence for efficient LLM processing. Extensive experiments demonstrate that BIMBA achieves state-of-the-art accuracy on multiple long-form VQA benchmarks, including PerceptionTest, NExT-QA, EgoSchema, VNBench, LongVideoBench, and Video-MME. Code, and models are publicly available at https://sites.google.com/view/bimba-mllm. |
| title | BIMBA: Selective-Scan Compression for Long-Range Video Question Answering |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2503.09590 |