See What You Need: Query-Aware Visual Intelligence through Reasoning-Perception Loops

Baoyun Peng; Lin Liu; Xiaodong Wang; Xinxin Dong; Yufei Wang; Yunlong Cao; Zixuan Dong

REVIEW 4 major objections 4 minor 44 references

See What You Need: Query-Aware Visual Intelligence through Reasoning-Perception Loops

T0 review · 4 major / 4 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A closed-loop reasoning-perception cycle, in which an LLM tells a VLM what to look at next, improves long-form video question answering accuracy by up to 6.9 points over static pipelines.

desk verdict Plausible method, unproven headline: the loop idea is sensible but the experiments never isolate it from backbone strength or compute budget. read the letter →

arxiv 2508.17932 v1 pith:64LRRNBX submitted 2025-08-25 cs.CV cs.AI

Zixuan Dong , Baoyun Peng , Yufei Wang , Lin Liu , Xinxin Dong , Yunlong Cao , Xiaodong Wang This is my paper · ORCID

classification cs.CVcs.AI

keywords videoquestionansweringclosed-loopreasoningreasoning-perceptioncoordinationiterativerefinementlong-formunderstandingtraining-freemultimodalpromptingkeyframelocalization

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Long-form video question answering usually runs perception first and reasoning second: a vision model absorbs the video into captions or frame features, and an LLM answers from that fixed summary. The paper argues this decoupling is the core failure mode, because different questions need different visual evidence from the same footage. It presents CAVIA, a training-free framework that closes the loop: the LLM states what information it lacks, the vision model extracts precisely that, and the cycle repeats until the LLM is confident. On EgoSchema, NExT-QA, and IntentQA the closed loop outperforms open-loop baselines by +5.3, +2.6, and +6.9 points, which the paper takes as evidence that reasoning-perception coordination is a scalable paradigm.

What carries the argument

The load-bearing object is the closed-loop reasoning-perception cycle itself: an LLM reasons about the query, identifies a specific information gap, emits a targeted prompt (e.g., "describe the tool manipulation sequence in frames 45–52"), a VLM extracts that visual evidence, and the new information is folded back into the LLM's reasoning until a confidence metric signals convergence. The three named components—hierarchical reasoning-guided localization, cross-modal semantic bridging, and confidence-driven iterative synthesis—are the mechanisms that make this loop precise rather than a generic caption-correction pass.

What would settle it

Run the same LLM, same VLM, same prompt templates, and same caption budget with the loop disabled—so the VLM extracts in one fixed pass and the LLM answers from that static evidence—and compare accuracy on EgoSchema, NExT-QA, and IntentQA; if the static pass matches or beats CAVIA, the central claim is false.

Watch

Extended reading notes

Core claim

The paper's central claim is that a closed reasoning-perception loop—where an LLM's reasoning gaps determine what a VLM extracts next—improves long-form video understanding without any training. The discovery is operationalized in CAVIA's three components: hierarchical reasoning-guided localization narrows from caption clusters to precise frames; cross-modal semantic bridging turns each identified reasoning gap into a targeted spatial-temporal extraction prompt; confidence-driven iterative synthesis repeats this cycle until the answer stabilizes. The paper reports state-of-the-art results on EgoSchema (65.7%), NExT-QA (76.1%), and IntentQA (73.8%), and interprets the gains on temporal and causal question types as direct evidence that query-driven visual extraction recovers information that static pipelines miss.

Load-bearing premise

The load-bearing premise is that the reported gains come from the closed-loop architecture rather than from the choice of LLM backbone, prompting, or compute; the manuscript's own Limitations section (which begins after the conclusion but is cut off in this excerpt) acknowledges unresolved issues, and the main comparison table mixes GPT-4 and GPT-4.1 results, selecting the best per benchmark in the abstract.

Editorial extensions

If this is right

Systems that currently caption the whole video first can be restructured so that the query drives what is looked at, without retraining the underlying LLM or VLM.
Iterative refinement with targeted prompts should be expected to produce the largest gains on temporal and causal question types, where the needed evidence is dispersed across frames.
The confidence-driven stopping rule gives a free efficiency knob: a deployment can trade compute for accuracy, since the loop halts once the answer is judged stable.
The same closed-loop design should transfer to other multimodal tasks that pair a reasoner with a perception module, such as temporal grounding or embodied decision-making, where evidence needs vary by query.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

An implication the paper leaves implicit is that the architecture's benefit can only be isolated if the same LLM, VLM, prompt templates, and compute budget are run with the loop open and closed; without that controlled experiment, the reported gains may overstate the architecture's contribution.
The excerpt stops at the opening line of the paper's own Limitations section (“While CAVIA demonstrates strong performance... several limitations”); reading the full list would clarify what the authors themselves consider unresolved, and that list should be weighed before accepting the paradigm claim.
A testable extension beyond the paper: measure whether the number of refinement iterations predicts answer correctness per question type; if it does, the confidence score could serve as a reliability oracle for deferring hard questions to a stronger model.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The core idea here is sound: instead of captioning the whole video once, let the LLM repeatedly ask the VLM for the specific frames and details it needs. That is a genuine extension of agentic video QA, and the combination of hierarchical localization, semantic bridging, and confidence-based stopping is new relative to DrVideo and VideoTree. The paper is clearly written and the system is described well enough to reconstruct. I believe the authors are onto something real.

The problem is the evidence. The abstract's state-of-the-art claim mixes GPT-4 and GPT-4.1 results, choosing whichever version gives the best number per benchmark. That alone would make me suspicious. The only same-backbone comparison is CAVIA GPT-4 vs VideoTree GPT-4, and while those gains are positive, the paper never shows that the closed loop is what causes them. The ablation in Figure 3 contrasts text-only, video-only, and the full system, but omits a single-pass open-loop variant using the same hierarchical localization and the same final prompt. Figure 4 shows accuracy rising with iterations, but no comparison against an equally expensive static pipeline. So the central causal claim—that the reasoning-perception loop drives the gains—is untested. The efficiency analysis (Figure 5) is a nice attempt, but it only measures captions, not total compute or latency.

Other soft spots: the Limitations section is literally empty, terminating at "several limitations" with no content. That is unprofessional and, combined with no code or prompt templates, makes verification hard. The trend line in Figure 4 is fit to four points and reported with r=0.932, which is not meaningful without error bars or more samples.

The benchmark gains against prior SOTA are computed against different backbones and prompt designs, so they are not apples-to-apples. I do not think the paper is dishonest—the method is described transparently, and the related work is fair. But the headline overclaims what is actually supported.

For a reader: this is a useful example of a promising architecture whose evaluation falls short of its claims. With a compute-matched open-loop control and consistent backbone across comparisons, the paper could be solid. Without that, the numbers don't prove the loop matters. I would not cite the SOTA numbers, but I might cite the method in a related-work survey.

Recommendation: send it to peer review, but with a clear request for a controlled ablation and de-cherry-picked abstract. A serious referee can make the paper address the confound. Desk-rejecting would waste a plausible idea; accepting as is would let a shaky causal claim into the literature.

Referee Report

4 major / 4 minor

Summary. The paper introduces CAVIA, a training-free closed-loop video question answering framework that iterates between an LLM reasoner and a VLM perceiver. It claims state-of-the-art results on EgoSchema, NExT-QA, and IntentQA through hierarchical reasoning-guided localization, cross-modal semantic bridging, and confidence-driven iterative synthesis. The manuscript includes comparisons with prior systems, ablations over modalities and refinement iterations, and an efficiency analysis.

Significance. If the closed-loop mechanism were shown to be the cause of the reported gains, the work would be a useful contribution to long-form video QA: it is training-free, integrates off-the-shelf models, and addresses a real limitation of static pipelines. The paper also makes an effort to study the effect of the number of refinement iterations and the number of captions. However, because the current experiments do not isolate the loop from backbone choice or compute budget, the central claim of a scalable paradigm is not yet substantiated.

major comments (4)

[Abstract and Table 1] The headline numbers in the abstract are not produced by a single system configuration: EgoSchema uses CAVIA-GPT-4 (65.7%), while NExT-QA (76.1%) and IntentQA (73.8%) use CAVIA-GPT-4.1. Prior SOTA baselines are GPT-3.5 or GPT-4, so the '+5.3%, +2.6%, +6.9%' gains are confounded by the choice of a stronger, never-compared backbone (GPT-4.1). The only same-backbone comparison, CAVIA-GPT-4 vs. VideoTree-GPT-4, shows smaller gains and a loss on Descriptive questions (79.2 vs. 81.3). The abstract should report a single configuration with matched baselines, and the paper should add GPT-4.1 baselines for the prior methods.
[§4.2.2 and Figure 3] The ablation that is meant to validate the closed-loop design compares text-only, video-only, and the full CAVIA pipeline, but it does not include the crucial open-loop control: a single-pass variant that uses the same hierarchical localization and the same final reasoning prompt but skips the iterative feedback. Without this control, the observed gains could be due to the larger number of captions or the longer reasoning trace rather than to the feedback loop itself. The same concern applies to Figure 4, which plots accuracy versus refinement iterations but does not compare against an equally expensive static pipeline.
[Limitations section] The manuscript's 'Limitations' section contains only the sentence fragment 'While CA VIA demonstrates strong performance across multiple benchmarks, several limitations' and then ends; it does not state any limitations. This is a missing statement of limitations, and it matters because the method has evident limitations (dependence on the backbone LLM, the number of iterations as a free parameter, API cost, and unknown generalization to other video domains). In addition, no code, prompts, or evaluation scripts are provided, so the reported numbers cannot be independently verified.
[Table 1 and Figures 3-5] No standard deviations, confidence intervals, or significance tests are reported for any of the main results (Table 1, Figures 3-5). The trend line in Figure 4 is fitted to only four points (refinement iterations 1-4); the reported r=0.932 is not statistically meaningful for such a small sample and should not be used to claim a monotone 'consistent improvement pattern.'

minor comments (4)

[Throughout] The method name is typeset inconsistently as 'CAVIA' in the abstract and 'CA VIA' throughout the body; please unify the notation.
[Table 1] Table 1 would be more informative if it reported the number of LLM/VLM calls or captions used by each method, since the central comparison is about efficiency as well as accuracy.
[Figure 5] Figure 5 compares accuracy versus number of captions, but the figure caption does not specify whether the LLoVi and VideoTree points are taken from the original papers or re-run under the same conditions; this should be stated.
[Abstract] The abstract states 'state-of-the-art performance' with the three benchmark numbers, but because these numbers come from two different configurations, the claim should be qualified.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: CAVIA's accuracy claims are externally measured on fixed benchmarks, with no fitted parameter renamed as a prediction and no load-bearing self-citation chain.

full rationale

The paper's central claim is an empirical performance result (EgoSchema, NExT-QA, IntentQA) for a training-free pipeline combining pretrained LLMs and VLMs. There is no mathematical derivation in which an output is defined as its input: the method's components (hierarchical reasoning-guided localization, cross-modal semantic bridging, confidence-driven iterative synthesis) are algorithmic procedures, and the reported accuracies come from benchmark evaluation, not from construction. No fitted parameter is estimated on a subset and then 'predicted' on a closely related quantity; no uniqueness theorem from the authors' prior work is invoked to force a choice; and no ansatz is imported solely through a self-citation. The paper does cite prior work including VideoTree, which appears to be from an overlapping group, but only as a baseline in Table 1 and as related work; the comparison is empirical rather than a load-bearing citation. Concerns that the SOTA comparisons are uncontrolled across backbone models (GPT-3.5 vs GPT-4 vs GPT-4.1), that the abstract mixes the best configuration per dataset, and that no same-backbone open-loop control is provided are threats to causal attribution and experimental rigor, not circularity under the definition used here. A non-finding is therefore appropriate.

Assumptions & free parameters 4 free parameters · 3 assumptions · 0 invented entities

The framework introduces no new entities. Its central claim rests on domain assumptions about LLM reasoning, VLM fidelity, and confidence calibration, plus several free parameters that are not fixed a priori. The reportable numbers also depend on the post-hoc choice of GPT-4 vs GPT-4.1 per benchmark.

free parameters (4)

Number of refinement iterations = 4
Figure 4 shows accuracy monotonically increasing to iteration 4; the stopping point is chosen by the LLM's confidence, but the fixed cap of 4 is a design choice not justified by data.
Backbone LLM version per benchmark = GPT-4 for EgoSchema, GPT-4.1 for NExT-QA and IntentQA
The abstract selects the best result across two GPT versions, indicating post-hoc selection rather than a fixed configuration.
Confidence threshold for stopping = not specified
The convergence criterion is described as confidence-driven but no threshold or calibration is reported.
Number of captions or segments = varies
The efficiency analysis varies the number of captions, but the exact sampling budget for each benchmark is not specified.

assumptions (3)

domain assumption The LLM (GPT-4 or GPT-4.1) can reliably identify which visual information is missing for answering a given question.
The entire loop depends on the LLM generating useful targeted prompts from information gaps; no evidence is provided beyond anecdotal examples.
domain assumption The VLM (unspecified) returns accurate, query-specific visual descriptions when prompted with frame ranges.
Cross-modal semantic bridging assumes the VLM can follow spatial-temporal directives without hallucination; no VLM error analysis is presented.
domain assumption The LLM's self-reported confidence correlates with answer correctness.
Confidence-driven iterative synthesis relies on this correlation; Figure 4 suggests the opposite, since accuracy keeps rising after earlier stops, indicating the confidence signal is not calibrated.

how reviews work

0 comments

Cite this review

Pith. "Pith review of See What You Need: Query-Aware Visual Intelligence through Reasoning-Perception Loops." pith.science (2026). https://pith.science/paper/64LRRNBX

@misc{pith2026250817932,
  author       = {Pith},
  title        = {Pith review of: See What You Need: Query-Aware Visual Intelligence through Reasoning-Perception Loops},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/64LRRNBX}},
  note         = {Machine review of arXiv:2508.17932}
}

read the original abstract

Human video comprehension demonstrates dynamic coordination between reasoning and visual attention, adaptively focusing on query-relevant details. However, current long-form video question answering systems employ rigid pipelines that decouple reasoning from perception, leading to either information loss through premature visual abstraction or computational inefficiency through exhaustive processing. The core limitation lies in the inability to adapt visual extraction to specific reasoning requirements, different queries demand fundamentally different visual evidence from the same video content. In this work, we present CAVIA, a training-free framework that revolutionizes video understanding through reasoning, perception coordination. Unlike conventional approaches where visual processing operates independently of reasoning, CAVIA creates a closed-loop system where reasoning continuously guides visual extraction based on identified information gaps. CAVIA introduces three innovations: (1) hierarchical reasoning, guided localization to precise frames; (2) cross-modal semantic bridging for targeted extraction; (3) confidence-driven iterative synthesis. CAVIA achieves state-of-the-art performance on challenging benchmarks: EgoSchema (65.7%, +5.3%), NExT-QA (76.1%, +2.6%), and IntentQA (73.8%, +6.9%), demonstrating that dynamic reasoning-perception coordination provides a scalable paradigm for video understanding.

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 26 canonical work pages

[1]

Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, and 1 others. 2025. Qwen2. 5-vl technical report. arXiv preprint arXiv:2502.13923

arXiv 2025
[2]

Ziyi Bai, Ruiping Wang, and Xilin Chen. 2023. Glance and focus: Memory prompting for multi-event video question answering. In Advances in Neural Information Processing Systems, volume 36

work page 2023
[3]

Ivana Bala z evi \'c , Yuge Shi, Pinelopi Papalampidi, Rahma Chaabouni, Skanda Koppula, and Olivier J H \'e naff. 2024. Memory consolidation enables long-context video understanding. arXiv preprint arXiv:2402.05861

arXiv 2024
[4]

Zhengxuan Chen, Shuo Wang, Deyang Yan, and Yushi Li. 2024. A spatio-temporl deepfake video detection method based on timesformer-cnn. In 2024 Third International Conference on Distributed Computing and Electrical Circuits and Electronics (ICDCECE), pages 1--6. IEEE

work page 2024
[5]

Xinxin Dong, Baoyun Peng, Haokai Ma, Yufei Wang, Zixuan Dong, Fei Hu, and Xiaodong Wang. 2025. https://arxiv.org/abs/2507.14784 Leadqa: Llm-driven context-aware temporal grounding for video question answering . Preprint, arXiv:2507.14784

arXiv 2025
[6]

Hao Fei, Shengqiong Wu, Wei Ji, Hanwang Zhang, Meishan Zhang, Mong-Li Lee, and Wynne Hsu. 2024. Video-of-thought: Step-by-step video reasoning from perception to cognition. In International Conference on Machine Learning (ICML)

work page 2024
[7]

Difei Gao, Luowei Zhou, Lei Ji, Linchao Zhu, Yi Yang, and Mike Zheng Shou. 2023. Mist: Multi-modal iterative spatial-temporal transformer for long-form video question answering. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14773--14783

2023
[8]

Sun Jianqiao, Su Yudi, Zhang Hao, Cheng Ziheng, Zeng Zequn, Wang Zhengjue, Chen Chunhui Qu, Bo, and Yuan Xin. 2025. https://nv.opticsjournal.net/J/AI/Issue/2025/1.html Snapcap: efficient snapshot compressive scene captioning . Advanced Imaging, 2(1):011003. Accessed: 2025-07-22

work page 2025

Show all 44 references

[9]

Yizhang Jin, Jian Li, Yexin Liu, Tianjun Gu, Kai Wu, Zhengkai Jiang, Muyang He, Bo Zhao, Xin Tan, Zhenye Gan, and 1 others. 2024. Efficient multimodal large language models: A survey. arXiv preprint arXiv:2405.10739

2024
[10]

Dohwan Ko, Ji Lee, Woo-Young Kang, Byungseok Roh, and Hyunwoo Kim. 2023 a . Large language models are temporal and causal reasoners for video question answering. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 4300--4316

2023
[11]

Dohwan Ko, Ji Lee, Woo-Young Kang, Byungseok Roh, and Hyunwoo Kim. 2023 b . Large language models are temporal and causal reasoners for video question answering. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 4300--4316

2023
[12]

Jiapeng Li, Ping Wei, Wenjuan Han, and Lifeng Fan. 2023 a . Intentqa: Context-aware video intent reasoning. In Proceedings of the IEEE/CVF international conference on computer vision, pages 11963--11974

2023
[13]

Yicong Li, Junbin Xiao, Chun Feng, Xiang Wang, and Tat-Seng Chua. 2023 b . Discovering spatio-temporal rationales for video question answering. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 13869--13878

2023
[14]

Ruotong Liao, Max Erler, Huiyu Wang, Guangyao Zhai, Gengyuan Zhang, Yunpu Ma, and Volker Tresp. 2024. Videoinsta: Zero-shot long video understanding via informative spatial-temporal reasoning with llms. arXiv preprint arXiv:2409.20365

2024 arXiv
[15]

Aixin Liu, Bei Feng, Bing Xue, Bingxuan Wang, Bochao Wu, Chengda Lu, Chenggang Zhao, Chengqi Deng, Chenyu Zhang, Chong Ruan, and 1 others. 2025. Deepseek-v3 technical report. arXiv preprint arXiv:2412.19437

2025 arXiv
[16]

Jie Ma, Pinghui Wang, Dechen Kong, Zewei Wang, Jun Liu, Hongbin Pei, and Junzhou Zhao. 2024. https://doi.org/10.1109/TPAMI.2024.3366154 Robust visual question answering: Datasets, methods, and future challenges . IEEE Transactions on Pattern Analysis and Machine Intelligence, ...

2024
[17]

Ziyu Ma, Chenhui Gou, Hengcan Shi, Bin Sun, Shutao Li, Hamid Rezatofighi, and Jianfei Cai. 2025. Drvideo: Document retrieval based long video understanding. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 18936--18946

2025
[18]

Karttikeya Mangalam, Raiymbek Akshulakov, and Jitendra Malik. 2023. Egoschema: A diagnostic benchmark for very long-form video language understanding. Advances in Neural Information Processing Systems, 36:46212--46244

2023
[19]

Meta AI . 2025. The llama 4 herd: The beginning of a new era of natively multimodal ai innovation. https://ai.meta.com/blog/llama-4-multimodal-intelligence/ . Accessed: 2025-04-05

2025
[20]

Juhong Min, Shyamal Buch, Arsha Nagrani, Minsu Cho, and Cordelia Schmid. 2024. Morevqa: Exploring modular reasoning models for video question answering. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 13235--13245

2024
[21]

Thong Thanh Nguyen, Zhiyuan Hu, Xiaobao Wu, Cong-Duy T Nguyen, See-Kiong Ng, and Anh Tuan Luu. 2024. Encoding and controlling global semantics for long-form video question answering. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pag...

2024
[22]

Pinelopi Papalampidi, Skanda Koppula, Shreya Pathak, Justin Chiu, Joe Heyward, Viorica Patraucean, Jiajun Shen, Antoine Miech, Andrew Zisserman, and Aida Nematzdeh. 2024. A simple recipe for contrastively pre-training video-first encoders beyond 16 frames. In Proceedings of th...

2024
[23]

Kanchana Ranasinghe, Xiang Li, Kumara Kahatapitiya, and Michael S Ryoo. 2025. Understanding long videos in one multimodal language model pass. arXiv preprint arXiv:2403.16998, 3

2025 arXiv
[24]

Saul Santos and 1 others. 2025. -video: A training-free approach to long video understanding via continuous-time memory consolidation. arXiv preprint arXiv:2501.19098

2025 arXiv
[25]

Himanshu Sharma and Anand Singh Jalal. 2022. Convolutional neural networks-based vqa model. In Proceedings of International Conference on Frontiers in Computing and Systems: COMSYS 2021, pages 109--116. Springer

2022
[26]

Yaojie Shen, Xin Gu, Kai Xu, Heng Fan, Longyin Wen, and Libo Zhang. 2023. Accurate and fast compressed video captioning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 15558--15567

2023
[27]

Enxin Song, Wenhao Chai, Guanhong Wang, Yucheng Zhang, Haoyang Zhou, Feiyang Wu, Haozhe Chi, Xun Guo, Tian Ye, Yanting Zhang, and 1 others. 2024. Moviechat: From dense token to sparse memory for long video understanding. In Proceedings of the IEEE/CVF Conference on Computer Vi...

2024
[28]

Raja Vavekanand and Kira Sam. 2024. Llama 3.1: An in-depth analysis of the next-generation large language model

2024
[29]

Xiaohan Wang, Yuhui Zhang, Orr Zohar, and Serena Yeung-Levy. 2025 a . Videoagent: Long-form video understanding with large language model as agent. In Computer Vision -- ECCV 2024, pages 58--76, Cham. Springer Nature Switzerland

2025
[30]

Yi Wang, Kunchang Li, Xinhao Li, Jiashuo Yu, Yinan He, Guo Chen, Baoqi Pei, Rongkun Zheng, Zun Wang, Yansong Shi, Tianxiang Jiang, Songze Li, Jilan Xu, Hongjie Zhang, Yifei Huang, Yu Qiao, Yali Wang, and Limin Wang. 2024 a . https://doi.org/10.1007/978-3-031-73013-9_23 Internv...

2024 doi
[31]

Yueqian Wang, Yuxuan Wang, Kai Chen, and Dongyan Zhao. 2024 b . Stair: Spatial-temporal reasoning with auditable intermediate results for video question answering. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 19215--19223

2024
[32]

Yufei Wang, Baoyun Peng, Zixuan Dong, Jia Fu, Xinxin Dong, Fei Hu, and Xiaodong Wang. 2025 b . Optigqa: Llm-driven query optimization for efficient visual grounding in adaptive video question answering. In Advanced Intelligent Computing Technology and Applications, pages 200--...

2025
[33]

Ziyang Wang, Shoubin Yu, Elias Stengel-Eskin, Jaehong Yoon, Feng Cheng, Gedas Bertasius, and Mohit Bansal. 2025 c . Videotree: Adaptive tree-based video representation for llm reasoning on long videos. In Proceedings of the Computer Vision and Pattern Recognition Conference, p...

2025
[34]

Haoning Wu, Dongxu Li, Bei Chen, and Junnan Li. 2024. Longvideobench: A benchmark for long-context interleaved video-language understanding. In Advances in Neural Information Processing Systems, volume 37

2024
[35]

Junbin Xiao, Xindi Shang, Angela Yao, and Tat-Seng Chua. 2021. Next-qa: Next phase of question-answering to explaining temporal actions. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 9777--9786

2021
[36]

Junbin Xiao, Pan Zhou, Tat-Seng Chua, and Shuicheng Yan. 2022. Video graph transformer for video question answering. In European Conference on Computer Vision, pages 39--58. Springer

2022
[37]

Geunhyuk Youk, Jihyong Oh, and Munchurl Kim. 2024. Fma-net: Flow-guided dynamic filtering and iterative feature refinement with multi-attention for joint video super-resolution and deblurring. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition...

2024
[38]

Shoubin Yu, Jaemin Cho, Prateek Yadav, and Mohit Bansal. 2023 a . Self-chained image-language model for video localization and question answering. In Advances in Neural Information Processing Systems, volume 36

2023
[39]

Shoubin Yu, Jaemin Cho, Prateek Yadav, and Mohit Bansal. 2023 b . Self-chained image-language model for video localization and question answering. Advances in Neural Information Processing Systems, 36:76749--76771

2023
[40]

Ce Zhang, Taixi Lu, Md Mohaiminul Islam, Ziyang Wang, Shoubin Yu, Mohit Bansal, and Gedas Bertasius. 2024 a . A simple llm framework for long-range video question-answering. Proceedings of the Conference on Empirical Methods in Natural Language Processing, pages 21715--21737

2024
[41]

Haoji Zhang, Yiqin Wang, Yansong Tang, Yong Liu, Jiashi Feng, Jifeng Dai, and Xiaojie Jin. 2024 b . Flash-vstream: Memory-based real-time understanding for long video streams. arXiv preprint arXiv:2406.08085

2024 arXiv
[42]

Junyang Zhang, Mu Yuan, Ruiguang Zhong, Puhan Luo, Huiyou Zhan, Ningkang Zhang, Chengchen Hu, and Xiang-Yang Li. 2025. A-vl: Adaptive attention for large vision-language models. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 22461--22469

2025
[43]

online" 'onlinestring :=

ENTRY address archivePrefix author booktitle chapter edition editor eid eprint eprinttype howpublished institution journal key month note number organization pages publisher school series title type volume year doi pubmed url lastchecked label extra.label sort.label short.list...
[44]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.

[1] [1]

Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, and 1 others. 2025. Qwen2. 5-vl technical report. arXiv preprint arXiv:2502.13923

arXiv 2025

[2] [2]

Ziyi Bai, Ruiping Wang, and Xilin Chen. 2023. Glance and focus: Memory prompting for multi-event video question answering. In Advances in Neural Information Processing Systems, volume 36

work page 2023

[3] [3]

Ivana Bala z evi \'c , Yuge Shi, Pinelopi Papalampidi, Rahma Chaabouni, Skanda Koppula, and Olivier J H \'e naff. 2024. Memory consolidation enables long-context video understanding. arXiv preprint arXiv:2402.05861

arXiv 2024

[4] [4]

Zhengxuan Chen, Shuo Wang, Deyang Yan, and Yushi Li. 2024. A spatio-temporl deepfake video detection method based on timesformer-cnn. In 2024 Third International Conference on Distributed Computing and Electrical Circuits and Electronics (ICDCECE), pages 1--6. IEEE

work page 2024

[5] [5]

Xinxin Dong, Baoyun Peng, Haokai Ma, Yufei Wang, Zixuan Dong, Fei Hu, and Xiaodong Wang. 2025. https://arxiv.org/abs/2507.14784 Leadqa: Llm-driven context-aware temporal grounding for video question answering . Preprint, arXiv:2507.14784

arXiv 2025

[6] [6]

Hao Fei, Shengqiong Wu, Wei Ji, Hanwang Zhang, Meishan Zhang, Mong-Li Lee, and Wynne Hsu. 2024. Video-of-thought: Step-by-step video reasoning from perception to cognition. In International Conference on Machine Learning (ICML)

work page 2024

[7] [7]

Difei Gao, Luowei Zhou, Lei Ji, Linchao Zhu, Yi Yang, and Mike Zheng Shou. 2023. Mist: Multi-modal iterative spatial-temporal transformer for long-form video question answering. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14773--14783

2023

[8] [8]

Sun Jianqiao, Su Yudi, Zhang Hao, Cheng Ziheng, Zeng Zequn, Wang Zhengjue, Chen Chunhui Qu, Bo, and Yuan Xin. 2025. https://nv.opticsjournal.net/J/AI/Issue/2025/1.html Snapcap: efficient snapshot compressive scene captioning . Advanced Imaging, 2(1):011003. Accessed: 2025-07-22

work page 2025