Static or Dynamic: Towards Query-Adaptive Token Selection for Video Question Answering

• URL: http://arxiv.org/abs/2504.21403v2
• Date: Mon, 15 Sep 2025 07:48:50 GMT
• Title: Static or Dynamic: Towards Query-Adaptive Token Selection for Video Question Answering
• Authors: Yumeng Shi, Quanyu Long, Wenya Wang,
• Abstract summary: Large volume of tokens generated from long videos presents challenges to memory efficiency and model performance.<n>We propose a novel token selection strategy, scexplore-then-text-select, that adaptively adjusts static and dynamic information based on question requirements.<n>Our framework is plug-and-play and can be seamlessly integrated within diverse video language models.
• Score: 21.906304473766056
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Video question answering benefits from the rich information in videos, enabling various applications. However, the large volume of tokens generated from long videos presents challenges to memory efficiency and model performance. To alleviate this, existing works propose to compress video inputs, but often overlook the varying importance of static and dynamic information across different queries, leading to inefficient token usage within limited budgets. We propose a novel token selection strategy, \textsc{explore-then-select}, that adaptively adjusts static and dynamic information based on question requirements. Our framework first explores different token allocations between key frames, which preserve spatial details, and delta frames, which capture temporal changes. Then it employs a query-aware attention-based metric to select the optimal token combination without model updates. Our framework is plug-and-play and can be seamlessly integrated within diverse video language models. Extensive experiments show that our method achieves significant performance improvements (up to 5.8\%) on multiple video question answering benchmarks. Our code is available at https://github.com/ANDgate99/Explore-Then-Select .

Related papers

• FLoC: Facility Location-Based Efficient Visual Token Compression for Long Video Understanding [55.700832127331324]
  FLoC is an efficient visual token compression framework based on the facility location function.<n>Our method achieves remarkable efficiency gains by swiftly selecting a compact subset of tokens.<n>Our approach is training-free, model-agnostic, and query-agnostic, providing a versatile solution.
  arXiv  Detail & Related papers  (2025-10-31T17:29:39Z)
• From Frames to Clips: Efficient Key Clip Selection for Long-Form Video Understanding [43.82717677801915]
  Video Large Language Models (VLMs) have achieved remarkable results on a variety of vision language tasks.<n>Their practical use is limited by the "needle in a haystack" problem: the massive number of visual tokens produced from raw video frames exhausts the model's context window.<n>We show that extending selection from isolated key frames to key clips, which are short, temporally coherent segments, improves video understanding.
  arXiv  Detail & Related papers  (2025-10-02T17:43:01Z)
• LoViC: Efficient Long Video Generation with Context Compression [68.22069741704158]
  We introduce LoViC, a DiT-based framework trained on million-scale open-domain videos.<n>At the core of our approach is FlexFormer, an expressive autoencoder that jointly compresses video and text into unified latent representations.
  arXiv  Detail & Related papers  (2025-07-17T09:46:43Z)
• Q-Frame: Query-aware Frame Selection and Multi-Resolution Adaptation for Video-LLMs [13.306662159600677]
  We introduce video QFrame, a novel approach for adaptive frame selection and multi-temporal scaling.<n>Q-Frame employs a training-free, plug-and-play strategy generated by a text-image matching network like CLIP.<n>We demonstrate Q-Frame's effectiveness through extensive experiments on benchmark datasets.
  arXiv  Detail & Related papers  (2025-06-27T11:30:51Z)
• Multimodal Long Video Modeling Based on Temporal Dynamic Context [13.979661295432964]
  We propose a dynamic long video encoding method utilizing the temporal relationship between frames, named Temporal Dynamic Context (TDC)<n>We segment the video into semantically consistent scenes based on inter-frame similarities, then encode each frame into tokens using visual-audio encoders.<n>To handle extremely long videos, we propose a training-free chain-of-thought strategy that progressively extracts answers from multiple video segments.
  arXiv  Detail & Related papers  (2025-04-14T17:34:06Z)
• STOP: Integrated Spatial-Temporal Dynamic Prompting for Video Understanding [48.12128042470839]
  We propose an integrated Spatial-TempOral dynamic Prompting (STOP) model. It consists of two complementary modules, the intra-frame spatial prompting and inter-frame temporal prompting. STOP consistently achieves superior performance against state-of-the-art methods.
  arXiv  Detail & Related papers  (2025-03-20T09:16:20Z)
• Whats in a Video: Factorized Autoregressive Decoding for Online Dense Video Captioning [71.94122309290537]
  We propose an efficient, online approach to generate dense captions for videos. Our model uses a novel autoregressive factorized decoding architecture. Our approach shows excellent performance compared to both offline and online methods, and uses 20% less compute.
  arXiv  Detail & Related papers  (2024-11-22T02:46:44Z)
• Beyond Training: Dynamic Token Merging for Zero-Shot Video Understanding [11.211803499867639]
  We propose DYTO, a novel dynamic token merging framework for zero-shot video understanding.<n> DYTO integrates a hierarchical frame selection and a bipartite token merging strategy to dynamically cluster key frames and selectively compress token sequences.<n>Experiments demonstrate the effectiveness of DYTO, achieving superior performance compared to both fine-tuned and training-free methods.
  arXiv  Detail & Related papers  (2024-11-21T18:30:11Z)
• HAVANA: Hierarchical stochastic neighbor embedding for Accelerated Video ANnotAtions [59.71751978599567]
  This paper presents a novel annotation pipeline that uses pre-extracted features and dimensionality reduction to accelerate the temporal video annotation process. We demonstrate significant improvements in annotation effort compared to traditional linear methods, achieving more than a 10x reduction in clicks required for annotating over 12 hours of video.
  arXiv  Detail & Related papers  (2024-09-16T18:15:38Z)
• TAM-VT: Transformation-Aware Multi-scale Video Transformer for Segmentation and Tracking [33.75267864844047]
  Video Object (VOS) has emerged as an increasingly important problem with availability of larger datasets and more complex and realistic settings. We propose a novel, clip-based DETR-style encoder-decoder architecture, which focuses on systematically analyzing and addressing aforementioned challenges. Specifically, we propose a novel transformation-aware loss that focuses learning on portions of the video where an object undergoes significant deformations.
  arXiv  Detail & Related papers  (2023-12-13T21:02:03Z)
• Characterizing Video Question Answering with Sparsified Inputs [55.7455981156755]
  We characterize a task with different input sparsity and provide a tool for doing that. Specifically, we use a Gumbel-based learnable selection module to adaptively select the best inputs for the final task. From our experiments, we have observed only 5.2%-5.8% loss of performance with only 10% of video lengths.
  arXiv  Detail & Related papers  (2023-11-27T21:00:20Z)
• TESTA: Temporal-Spatial Token Aggregation for Long-form Video-Language Understanding [20.16000249533665]
  TESTA condenses video semantics by adaptively aggregating similar frames, as well as similar patches within each frame. Building upon TESTA, we introduce a pre-trained video-language model equipped with a divided space-time token aggregation module in each video block. We evaluate our model on five datasets for paragraph-to-video retrieval and long-form VideoQA tasks.
  arXiv  Detail & Related papers  (2023-10-29T16:25:32Z)
• Differentiable Resolution Compression and Alignment for Efficient Video Classification and Retrieval [16.497758750494537]
  We propose an efficient video representation network with Differentiable Resolution Compression and Alignment mechanism. We leverage a Differentiable Context-aware Compression Module to encode the saliency and non-saliency frame features. We introduce a new Resolution-Align Transformer Layer to capture global temporal correlations among frame features with different resolutions.
  arXiv  Detail & Related papers  (2023-09-15T05:31:53Z)
• MIST: Multi-modal Iterative Spatial-Temporal Transformer for Long-form Video Question Answering [73.61182342844639]
  We introduce a new model named Multi-modal Iterative Spatial-temporal Transformer (MIST) to better adapt pre-trained models for long-form VideoQA. MIST decomposes traditional dense spatial-temporal self-attention into cascaded segment and region selection modules. Visual concepts at different granularities are then processed efficiently through an attention module.
  arXiv  Detail & Related papers  (2022-12-19T15:05:40Z)
• Temporal Saliency Query Network for Efficient Video Recognition [82.52760040577864]
  Video recognition is a hot-spot research topic with the explosive growth of multimedia data on the Internet and mobile devices. Most existing methods select the salient frames without awareness of the class-specific saliency scores. We propose a novel Temporal Saliency Query (TSQ) mechanism, which introduces class-specific information to provide fine-grained cues for saliency measurement.
  arXiv  Detail & Related papers  (2022-07-21T09:23:34Z)
• Efficient Video Transformers with Spatial-Temporal Token Selection [68.27784654734396]
  We present STTS, a token selection framework that dynamically selects a few informative tokens in both temporal and spatial dimensions conditioned on input video samples. Our framework achieves similar results while requiring 20% less computation.
  arXiv  Detail & Related papers  (2021-11-23T00:35:58Z)
• EAN: Event Adaptive Network for Enhanced Action Recognition [66.81780707955852]
  We propose a unified action recognition framework to investigate the dynamic nature of video content. First, when extracting local cues, we generate the spatial-temporal kernels of dynamic-scale to adaptively fit the diverse events. Second, to accurately aggregate these cues into a global video representation, we propose to mine the interactions only among a few selected foreground objects by a Transformer.
  arXiv  Detail & Related papers  (2021-07-22T15:57:18Z)
• Dense-Caption Matching and Frame-Selection Gating for Temporal Localization in VideoQA [96.10612095576333]
  We propose a video question answering model which effectively integrates multi-modal input sources and finds the temporally relevant information to answer questions. Our model is also comprised of dual-level attention (word/object and frame level), multi-head self-cross-integration for different sources (video and dense captions), and which pass more relevant information to gates. We evaluate our model on the challenging TVQA dataset, where each of our model components provides significant gains, and our overall model outperforms the state-of-the-art by a large margin.
  arXiv  Detail & Related papers  (2020-05-13T16:35:27Z)
• Scene-Adaptive Video Frame Interpolation via Meta-Learning [54.87696619177496]
  We propose to adapt the model to each video by making use of additional information that is readily available at test time. We obtain significant performance gains with only a single gradient update without any additional parameters.
  arXiv  Detail & Related papers  (2020-04-02T02:46:44Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.