Pith. sign in

REVIEW 4 major objections 5 minor 60 references

VTO: Visual Tool Orchestration for Video Anomaly Detection

T0 review · 4 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read VTO claims that process-supervised reinforcement learning—rule-based rewards plus a vision-language judge—raises an 8B agent's whole-response accuracy on interrelated video-anomaly tool queries to 95.14%, a 10.2-point gain over SFT.

desk verdict A useful benchmark and a believable internal result, but the reward/evaluation loop is anchored to one model family's trajectories, so the headline gain partly measures alignment with Qwen2.5-VL rather than better anomaly detection. read the letter →

arxiv 2608.08219 v1 pith:4CXTBTJ4 submitted 2026-08-08 cs.CV cs.AI

Rui Wang , Yeteng Wu , Xianling Zhang , Mengshi Qi This is my paper · ORCID
classification cs.CVcs.AI
keywords videoanomalydetectiontoollearningmultimodalagentsprocess-supervisedreinforcementGRPOLLM-as-a-judgeorchestrationVAD-Toolbenchmark
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

VTO reframes video anomaly detection as a multi-step tool-orchestration problem: an agent must call several specialized vision models in sequence to trace how an incident unfolds. The paper's claim is that supervising each reasoning step rather than only the final answer fixes two failures of prior approaches—supervised fine-tuned agents imitate a fixed path, and outcome-based reinforcement learning agents stop too early. To do this, VTO combines exact rule-based rewards (tool name, parameters, completion, efficiency) with judgments from a large vision-language model scoring logicality, relevance, and completeness, then optimizes the policy with GRPO. On the new VAD-Tool benchmark, the framework reports 95.14% whole-response accuracy for interrelated tool queries on an 8B backbone, an absolute gain of 10.2 points over SFT and higher than a 72B model. If correct, this gives a practical recipe for making small multimodal agents orchestrate complex visual analyses without waiting for bigger models.

What carries the argument

The load-bearing mechanism is the dual-reward Process-Supervised Cognitive Alignment wrapped around Group Relative Policy Optimization (GRPO). At each step the total reward is $r_t = R_{\text{rule}} + R_{\text{LLM}}$: $R_{\text{rule}}$ gives exact-match credit for task completion, tool name, and tool parameters, plus an efficiency penalty for extra steps, while $R_{\text{LLM}}$ uses a 72B vision-language judge to score logicality, relevance, and completeness of the trajectory. Rewards are accumulated over the trajectory, normalized within a group of eight sampled rollouts to form advantages, and used in a clipped GRPO objective with a KL penalty to the SFT reference model. This mechanism is what explicitly penalizes premature termination: a trajectory that stops after the first anomaly detection receives a low completeness score at every subsequent step.

What would settle it

Have several independent security-video experts annotate the same 200 interrelated test queries with every tool sequence they consider correct, then re-score VTO accepting any expert-valid sequence instead of requiring exact match. If experts frequently propose more than one valid path, the exact-match reward will have marked legitimate alternatives as errors, and the reported 95.14% whole accuracy and the 10.2-point improvement over SFT would shrink under any-of-valid scoring.

Watch

Extended reading notes

Core claim

The paper argues that the bottleneck in video anomaly detection with tool-using agents is not the tools or the backbone but the reward granularity. Its central discovery, on its own terms, is that process-supervised cognitive alignment—dense step-level rewards combining deterministic rule checks with a foundation-model judge—converts an SFT-trained 8B model from a path mimicker into a reliable multi-step orchestrator. On interrelated queries in VAD-Tool, VTO reports perfect decision and tool-selection accuracy, 95.60% input accuracy, and 95.14% whole-response accuracy, against 89.74% for SFT and 67.40% for a 72B model used zero-shot. The ablations attribute the gain to every reward component: without the task-completion reward accuracy collapses to 29.92%, without parameter-match supervision it falls to 77.66%, and without the LLM judge it drops to 93.70%.

Load-bearing premise

The load-bearing premise is that every benchmark query has exactly one correct tool-use trajectory and that the human-checked Qwen-generated annotations plus the 72B judge's scores capture what good reasoning is; if alternative valid sequences exist or the judge is biased, the accuracy numbers measure agreement with the annotator rather than anomaly-detection skill.

Editorial extensions

If this is right

  • SFT alone leaves a multi-step tool-orchestration gap; adding process-supervised RL closes most of it, raising interrelated whole-response accuracy from 89.74% to 95.14% on Qwen3-VL-8B.
  • An 8B parameter agent trained with process rewards can outperform a 72B model used zero-shot or few-shot on interrelated anomaly queries.
  • Penalizing incomplete causal chains at each step prevents the premature-termination failure of outcome-based RL in cascading-hazard scenarios.
  • A hierarchy of 12 visual tools, from person re-identification to fire detection, provides a usable action space for explainable video anomaly detection.
  • Exact tool-name and parameter rewards anchor the agent to real visual observations, reducing hallucinated frame or coordinate grounding.

Reading between the lines

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

  • The paper leaves implicit that the same rule-plus-judge process reward could transfer to other multi-step visual reasoning settings, such as medical imaging or autonomous driving, where stopping after the first finding is also dangerous.
  • An unexamined consequence is that exact-match rewards train the agent toward the annotator's single path; if alternative tool sequences are legitimate, a softer any-acceptable-path reward would be needed, and the 10.2-point gap would likely shrink.
  • A testable extension is to compare the LLM judge's completeness scores against human expert ratings on held-out trajectories; large disagreement would indicate the policy is optimizing judge-pleasing style rather than causal completeness.
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.

Referee Report

4 major / 5 minor

Summary. The paper proposes VTO, a process-supervised reinforcement learning framework for visual tool orchestration in video anomaly detection (VAD). The method first supervised fine-tunes a vision-language model on ReAct-format trajectories, then applies GRPO with a dual reward composed of rule-based exact-match rewards and an LLM-as-a-judge reward. The authors also introduce VAD-Tool, a benchmark with 12 vision tools and over 120k instruction-trajectory pairs, and report that VTO improves whole-response accuracy over SFT, e.g., from 89.74% to 95.14% on interrelated-tool responses with a Qwen3-VL-8B backbone (Table 2).

Significance. If validated, VTO would be a useful contribution to agentic VAD: it provides a deliberately constructed multi-tool benchmark, a concrete two-stage training recipe, and evidence that process-level rewards can improve multi-step tool scheduling. The paper also ships code and data, which is a practical strength. However, the headline result is currently difficult to interpret because the reward and evaluation metrics are anchored to exactly one Qwen2.5-VL-generated trajectory per query, and the LLM judge is from the same model family, so the reported gains may reflect alignment with one annotator path rather than improved anomaly detection. The absence of an outcome-based RL baseline also leaves the paper's core motivation untested.

major comments (4)
  1. [Section 3.3, Section 4.3.1, Eq. (3), Eq. (7)] The evaluation and training are circular relative to the benchmark's ground truth. The ReAct trajectories are generated by Qwen2.5-VL (Section 3.3), the rule-based rewards in Eq. (3) reward exact matches to those same trajectories, the LLM judge in Eq. (4) is Qwen2.5-VL-72B (Section 4.3.1), and the Whole metric in Eq. (7) requires exact equality of every decision, tool name, and parameter string with the same Qwen2.5-VL-generated annotation. The human review described in Section 3.3 checks logical correctness but does not enumerate alternative valid tool sequences. Consequently, the reported 5.4-point Interrelated Whole improvement in Table 2 may measure how faithfully VTO reproduces one specific annotator path rather than how well it detects anomalies. I ask the authors to add an outcome-based evaluation (e.g., human-judged correctness of the final anomaly verdict, independent of the tool path) or a multi-annotation study showing that alternative valid trajectories are not penalized.
  2. [Section 4.1, Section 5.2, Table 2] The paper's central motivation is that standard outcome-based RL causes premature termination, but no outcome-based RL baseline is reported. Without a GRPO baseline using only a final outcome reward (or a coarse completion reward) under the same compute and rollout settings, the experiments cannot establish that process supervision rather than RL in general is responsible for the gains. The ablation in Table 3 removes individual VTO components but never compares against pure outcome-based RL. Please add such a baseline and report it in Table 2 or Table 3.
  3. [Section 4.3.1, Eq. (2)-(4)] The reward model is underspecified. The paper does not give the functional forms or numerical scales of Rcomp, Rtool, Rparam, Reff, Rlogic, Rrel, and Rcomplete, nor the combination weights in Eq. (2), nor the discount factor gamma used in the cumulative return. These are free parameters that directly determine the optimization landscape, and the ablation results in Table 3 cannot be reproduced or interpreted without them. Please provide the exact reward definitions, coefficient values, and gamma, or point to released code that contains them.
  4. [Table 2] No variance or significance information is reported for the main accuracy comparisons. The headline numbers (e.g., 89.74% vs. 95.14% for Interrelated Whole) come from a single table with no confidence intervals or multiple-seed results, so it is unclear whether the differences are stable. I request at least three seeds with mean and standard deviation for the main VTO and SFT comparisons, or an explicit statement of why a single run is sufficient.
minor comments (5)
  1. [Abstract and Table 2] The abstract's '10.2% absolute accuracy improvement' appears to refer to the All Responses Whole column (86.19% to 96.39%), while the interrelated-tool improvement is 5.4 points; please state explicitly which number is being reported to avoid confusion.
  2. [Section 3.3] The term 'human-in-the-loop' is used for a pipeline where Qwen2.5-VL generates the trajectories and human experts review them; the level of human intervention (e.g., fraction of trajectories edited or rejected) is not quantified, which would help assess the independence of the benchmark from the generator.
  3. [Figure 4 and Section 4.3] Figure 4 labels the reward phase as 'Step 3' while Section 4.3 describes Rollout, Reward Calculation, and Policy Optimization; the figure and text should use consistent step numbering.
  4. [Table 2] The 'Prompt' column uses 'Zero-shot' for VTO rows even though VTO is a trained model; consider renaming the column to 'Setting' or 'Paradigm' to avoid implying VTO is inference-only.
  5. [Throughout] The benchmark name is inconsistently hyphenated as 'VAD-Tool' and 'V AD-Tool'; please use a single spelling throughout.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: VTO's headline gain is an empirical, held-out result; the Qwen2.5-VL judge/generator overlap is a validity caveat, not a circular derivation.

full rationale

The paper's central claim is an empirical result on a held-out benchmark, not a formal derivation. The rule-based rewards (Eq. 3) and the Whole metric (Eq. 7) both use exact match against the same ground-truth annotations, but this is standard objective/evaluation alignment: the training and test splits are disjoint, since Section 3.4 states that the physical video scenes and linguistic semantics in the testing benchmark are strictly distinct from those in the training dataset. Thus test accuracy is not implied by construction. The use of Qwen2.5-VL-72B as LLM judge (Section 4.3.1) while Qwen2.5-VL generated the initial annotations (Section 3.3) is a legitimate validity caveat: the benchmark may reward agreement with one generator's preferred trajectory style and penalize alternative valid tool sequences. However, that is a correctness/external-validity concern, not a circular derivation: the reported 10.2% gain is not logically forced by the reward definition. The citations to the authors' own prior work (e.g., [55] for tool taxonomy) are not load-bearing for the headline result. Therefore no significant circularity is found.

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

The central claim rests on the assumption that the LLM-generated, expert-reviewed annotations define correct behavior, and that the reward and evaluation functions (exact match, LLM judge) capture orchestration quality. The unspecified reward coefficients and discount factor are additional free choices.

free parameters (3)
  • Efficiency penalty coefficient
    Reff applies a penalty proportional to the number of steps (Section 4.3.1, Eq. 3), but the proportionality constant is not specified, so it is a hand-chosen parameter affecting the reward.
  • Reward combination weights = implicitly equal
    Equation (2) sums R_rule and R_LLM directly without specified scaling or normalization, so the relative weighting is a free choice.
  • Discount factor gamma
    Used in the cumulative return R(tau) but no value is reported (Section 4.3.1).
assumptions (4)
  • domain assumption The 12 off-the-shelf visual tools are sufficient and appropriate for VAD tool orchestration.
    The entire benchmark is built on this tool taxonomy (Section 3.2, Table 1).
  • domain assumption Qwen2.5-VL-generated ReAct trajectories, after expert review, are correct ground truth.
    Section 3.3: initial trajectories are LLM-generated and human-reviewed; the rule-based rewards treat them as exact targets.
  • domain assumption Qwen2.5-VL-72B provides valid logicality, relevance, and completeness judgments.
    Section 4.3.1: the LLM judge's scalar scores are used as reward signals without validation against human preferences.
  • ad hoc to paper Exact match to a single annotated tool sequence is the right evaluation metric.
    Equation (7) requires every step to match the annotated component; alternative correct sequences receive zero credit. This is an ad hoc choice that may penalize valid multi-step solutions.

how reviews work

0 comments
Cite this review

Pith. "Pith review of VTO: Visual Tool Orchestration for Video Anomaly Detection." pith.science (2026). https://pith.science/paper/4CXTBTJ4

@misc{pith2026260808219,
  author       = {Pith},
  title        = {Pith review of: VTO: Visual Tool Orchestration for Video Anomaly Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4CXTBTJ4}},
  note         = {Machine review of arXiv:2608.08219}
}
read the original abstract

Video anomaly detection (VAD) is a critical yet challenging task due to the complex and diverse nature of real-world scenarios. Traditional deep learning approaches are fundamentally limited by poor generalization across diverse scenarios. While multimodal agents offer a promising tool-learning paradigm for VAD, current systems relying on supervised fine-tuning struggle with complex orchestration, and standard reinforcement learning often causes premature termination due to coarse-grained outcome rewards. To address these challenges, we propose VTO, a process-supervised reinforcement learning framework. Moving beyond static tool usage, VTO enables the agent to dynamically explore and interact with the environment. Specifically, we introduce a foundation model-driven cognitive evaluator to provide context-aware semantic feedback, which is seamlessly integrated into a Process-Supervised Cognitive Alignment that delivers fine-grained, step-wise supervision. By explicitly penalizing logical truncation and rewarding complete causal chains, the agent optimizes its multi-step reasoning policy for interrelated tool orchestration. To support our proposed framework, we meticulously crafted VAD-Tool, a hierarchical visual tool set comprising 12 specialized vision tools spanning from entity tracking to high-stakes hazard detection, and established the corresponding benchmark for rigorous multi-step reasoning evaluation. Extensive experiments on VAD-Tool demonstrate that VTO significantly outperforms baselines, achieving up to a 10.2\% absolute accuracy improvement in tool scheduling. Code and data are available at https://github.com/MICLAB-BUPT/VTO.

Figures

Figures reproduced from arXiv: 2608.08219 by the authors.

Figure 1
Figure 1. Comparison of VAD paradigms. (a) Traditional [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The inference pipeline of the proposed framework. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Overview of VAD-Tool dataset construction and [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: The overall pipeline of the VTO framework. Step 1: Supervised Fine-Tuning equips the base model with foundational [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Qualitative examples from the VAD-Tool benchmark illustrating structured reasoning trajectories across three diverse [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

60 extracted references · 48 canonical work pages

  1. [1]

    Shuai Bai, Yuxuan Cai, Ruizhe Chen, Keqin Chen, Xionghui Chen, Zesen Cheng, Lianghao Deng, Wei Ding, Chang Gao, Chunjiang Ge, et al . 2025. Qwen3-vl technical report.arXiv preprint arXiv:2511.21631(2025)

    arXiv 2025

  2. [2]

    Sijia Chen, Yibo Wang, Yi-Feng Wu, Zhao Xu, Weihua Luo, Kaifu Zhang, and Lijun Zhang. 2024. Advancing Tool-Augmented Large Language Models: Inte- grating Insights from Errors in Inference Trees. InAdvances in Neural Information Processing Systems (NeurIPS)

    work page 2024

  3. [3]

    Wei-Lin Chiang, Zhuohan Li, Ziqing Lin, Ying Sheng, Zhanghao Wu, Hao Zhang, Lianmin Zheng, Siyuan Zhuang, Yonghao Zhuang, Joseph E Gonzalez, et al. 2023. Vicuna: An open-source chatbot impressing gpt-4 with 90%* chatgpt quality.See https://vicuna. lmsys. org (accessed 14 April 2023)2, 3 (2023), 6

    2023

  4. [4]

    Wei Deng, Mengshi Qi, and Huadong Ma. 2025. Global-local tree search in vlms for 3d indoor scene generation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 8975–8984

    work page 2025

  5. [5]

    Chao Fan, Junjie Liang, Chuanfu Shen, et al . 2023. OpenGait: Revisiting Gait Recognition Towards Better Practicality. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

    work page 2023

  6. [6]

    Shibo Gao, Peng Yang, Hui Guo, Yuan Liu, Yifei Chen, Shuo Li, Haoran Zhu, Jiajun Xu, and Xiaoyun Zhang. 2025. The Evolution of Video Anomaly Detection: A Unified Framework from DNN to MLLM.arXiv preprint arXiv:2507.21649(2025)

    arXiv 2025

  7. [7]

    Zhi Gao, Bofei Zhang, Pengxiang Li, Xiaojian Ma, Tao Yuan, Yue Fan, Yuwei Wu, Yunde Jia, Song-Chun Zhu, and Qing Li. 2025. Multi-modal agent tuning: Building a vlm-driven agent for efficient tool usage. InInternational Conference on Learning Representations (ICLR)

    work page 2025

  8. [8]

    Dong Gong, Lingqiao Liu, Vuong Le, Budhaditya Saha, Moussa Reda Mansour, Svetha Venkatesh, and Anton van den Hengel. 2019. Memorizing normality to detect anomaly: Memory-augmented deep autoencoder for unsupervised anomaly detection. InProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)

    work page 2019

Show all 60 references
  1. [9]

    Mahmudul Hasan, Jonghyun Choi, Jan Neumann, Amit K Roy-Chowdhury, and Larry S Davis. 2016. Learning temporal regularity in video sequences. InProceed- ings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

    2016

  2. [10]

    Lingxiao He, Xingyu Liao, Wu Liu, et al. 2023. FastReID: A Pytorch Toolbox for General Instance Re-identification. InProceedings of the 31st ACM International Conference on Multimedia (ACMMM)

    2023

  3. [11]

    Mohammad Mahdi Hemmatyar, Mohammad Jafari, Mohammad Amin Yousefi, Mohammad Reza Nemati, Mahdi Azadani, and Hamid Reza Rastad. 2025. HyCo- VAD: A Hybrid SSL-LLM Model for Complex Video Anomaly Detection.arXiv preprint arXiv:2509.22544(2025)

    2025

  4. [12]

    Dingkang Liang, Wei Xu, and Xiang Bai. 2022. An end-to-end transformer model for crowd localization. InEuropean Conference on Computer Vision (ECCV)

    2022

  5. [13]

    Siwei Liu, Jinyuan Fang, Han Zhou, Yingxu Wang, and Zaiqiao Meng. 2025. SEW: Self-Evolving Agentic Workflows for Automated Code Generation.arXiv preprint arXiv:2505.18646(2025)

    2025 arXiv

  6. [14]

    Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Qing Jiang, Chunyuan Li, Jianwei Yang, Hang Su, et al. 2024. Grounding dino: Marry- ing dino with grounded pre-training for open-set object detection. InEuropean conference on computer vision (ECCV)

    2024

  7. [15]

    Xinchen Liu, Wu Liu, Huadong Ma, et al . 2016. Large-Scale Vehicle Re- Identification in Urban Surveillance Videos. InIEEE International Conference on Multimedia and Expo (ICME)

    2016

  8. [16]

    Yanming Liu, Xinyue Peng, Jiannan Cao, Yuwei Zhang, Xuhong Zhang, Sheng Cheng, Xun Wang, Jianwei Yin, and Tianyu Du. 2025. Tool-Planner: Task Plan- ning with Clusters across Multiple Tools. InProceedings of the International Conference on Learning Representations (ICLR)

    2025

  9. [17]

    Peng Lu, Tao Jiang, Yining Li, Xiangtai Li, Kai Chen, and Wenming Yang. 2024. Rtmo: Towards high-performance one-stage real-time multi-person pose estima- tion. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition (CVPR)

    2024

  10. [18]

    Changsheng Lv, Zijian Fu, and Mengshi Qi. 2026. Robo-SGG: Exploiting Layout- Oriented Normalization and Restitution Can Improve Robust Scene Graph Gen- eration. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). 39282–39292

    2026

  11. [19]

    Changsheng Lv, Mengshi Qi, Xia Li, Zhengyuan Yang, and Huadong Ma. 2024. Sgformer: Semantic graph transformer for point cloud-based 3d scene graph generation. InProceedings of the AAAI Conference on Artificial Intelligence, Vol. 38. 4035–4043

    2024

  12. [20]

    Changsheng Lv, Mengshi Qi, Liang Liu, and Huadong Ma. 2025. T2sg: Traffic topology scene graph for topology reasoning in autonomous driving. InProceed- ings of the Computer Vision and Pattern Recognition Conference. 17197–17206

    2025

  13. [21]

    Hui Lv and Qianru Sun. 2024. Video Anomaly Detection and Explanation via Large Language Models.arXiv preprint arXiv:2401.05702(2024)

    2024 arXiv

  14. [22]

    Chang Ma, Junlei Zhang, Zhihao Zhu, Cheng Yang, Yujiu Yang, Yaohui Jin, Zhen- zhong Lan, Lingpeng Kong, and Junxian He. 2024. Agentboard: An analytical evaluation board of multi-turn llm agents.Advances in neural information pro- cessing systems (NeurIPS)(2024)

    2024

  15. [23]

    Da Ma, Ziyue Yang, Hongshen Xu, Haotian Fang, Kai Yu, and Lu Chen. 2026. Empowering LLM Tool Invocation with Tool-call Reward Model. InProceedings of the International Conference on Learning Representations (ICLR)

    2026

  16. [24]

    Mengshi Qi, Changsheng Lv, and Huadong Ma. 2025. Robust disentangled counterfactual learning for physical audiovisual commonsense reasoning.IEEE Transactions on Pattern Analysis and Machine Intelligence(2025)

    2025

  17. [25]

    Mengshi Qi, Jiaxuan Peng, Jie Zhang, Juan Zhu, Yong Li, and Huadong Ma. 2026. Synergistic tensor and pipeline parallelism.Advances in Neural Information Processing Systems38 (2026), 157997–158022

    2026

  18. [26]

    Mengshi Qi, Jie Qin, Yi Yang, Yunhong Wang, and Jiebo Luo. 2021. Semantics- aware spatial-temporal binaries for cross-modal video retrieval.IEEE Transactions on Image Processing30 (2021), 2989–3004

    2021

  19. [27]

    Mengshi Qi, Yeteng Wu, Wulian Yun, Xianlin Zhang, and Huadong Ma. 2026. Explainable action form assessment by exploiting multimodal chain-of-thoughts reasoning.IEEE Transactions on Image Processing(2026)

    2026

  20. [28]

    Mengshi Qi, Hao Ye, Jiaxuan Peng, and Huadong Ma. 2025. Action quality assessment via hierarchical pose-guided multi-stage contrastive regression.IEEE Transactions on Image Processing(2025)

    2025

  21. [29]

    Mengshi Qi, Pengfei Zhu, Xiangtai Li, Xiaoyang Bi, Lu Qi, Huadong Ma, and Ming-Hsuan Yang. 2025. DC-SAM: In-Context Segment Anything in Images and Videos via Dual Consistency.IEEE Transactions on Pattern Analysis and Machine Intelligence(2025)

    2025

  22. [30]

    Zhihong Shao, Peiyi Wang, Qihao Zhu, Runxin Xu, Junxiao Song, Xiao Bi, Haowei Zhang, Mingchuan Zhang, YK Li, et al. 2024. Deepseekmath: Pushing the limits of mathematical reasoning in open language models.arXiv preprint arXiv:2402.03300 (2024)

    2024 arXiv

  23. [31]

    Noah Shinn, Federico Cassano, Ashwin Gopinath, Karthik Narasimhan, and Shunyu Yao. 2023. Reflexion: Language agents with verbal reinforcement learning. Advances in neural information processing systems (NeurIPS)(2023)

    2023

  24. [32]

    Waqas Sultani, Chen Chen, and Mubarak Shah. 2018. Real-world anomaly detec- tion in surveillance videos. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

    2018

  25. [33]

    Hugo Touvron, Louis Martin, Kevin Stone, et al. 2023. Llama 2: Open Foundation and Fine-Tuned Chat Models.arXiv preprint arXiv:2307.09288(2023)

    2023 arXiv

  26. [34]

    Rejin Varghese and M Sambath. 2024. Yolov8: A novel object detection algorithm with enhanced performance and robustness. In2024 International conference on advances in data engineering and intelligent computing systems (ADICS). IEEE, 1–6

    2024

  27. [35]

    Chien-Yao Wang, Alexey Bochkovskiy, and Hong-Yuan Mark Liao. 2023. YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors. InProceedings of the IEEE/CVF conference on computer vision and pattern recogni- tion. 7464–7475

    2023

  28. [36]

    Huacan Wang, Ronghao Chen, et al. 2025. SE-Agent: Self-Evolution Trajectory Optimization in Multi-Step Reasoning with LLM-Based Agents. InAdvances in Neural Information Processing Systems (NeurIPS)

    2025

  29. [37]

    Rui Wang, Mengshi Qi, Yingxia Shao, Anfu Zhou, and Huadong Ma. 2025. PITN: Physics-Informed Temporal Networks for Cuffless Blood Pressure Estimation. IEEE Transactions on Mobile Computing(2025)

    2025

  30. [38]

    Xingyao Wang et al . 2024. SELF: Language-Driven Self-Evolution for Large Language Model. InProceedings of the International Conference on Learning Rep- resentations (ICLR)

    2024

  31. [39]

    Zihan Wang, Kangrui Wang, et al. 2025. RAGEN: Understanding Self-Evolution in LLM Agents via Multi-Turn Reinforcement Learning.arXiv preprint arXiv:2504.20073(2025)

    2025 arXiv

  32. [40]

    Peng Wu, Xuerong Zhou, Guansong Pang, Lingru Zhou, Qingsen Yan, Peng Wang, and Yanning Zhang. 2024. Vadclip: Adapting vision-language models for weakly supervised video anomaly detection. InProceedings of the AAAI conference on artificial intelligence (AAAI)

    2024

  33. [41]

    Zhenbo Xu, Wei Yang, Ajin Meng, et al. 2018. Towards End-to-End License Plate Detection and Recognition: A Large Dataset and Baseline. InEuropean Conference on Computer Vision (ECCV)

    2018

  34. [42]

    Yuchen Yang, Kwonjoon Lee, Behzad Dariush, Yinzhi Cao, and Shao-Yuan Lo

  35. [43]

    Zhiwei Yang, Chen Gao, and Mike Zheng Shou. 2025. PANDA: Towards Gener- alist Video Anomaly Detection via Agentic AI Engineer. InAdvances in Neural Information Processing Systems (NeurIPS). MM ’26, November 10–14, 2026, Rio de Janeiro, Brazil Wang et al

    2025

  36. [44]

    Shunyu Yao, Jeffrey Zhao, Dian Yu, et al . 2023. ReAct: Synergizing Reason- ing and Acting in Language Models. InInternational Conference on Learning Representations (ICLR)

    2023

  37. [45]

    Hao Ye, Mengshi Qi, Zhaohong Liu, Liang Liu, and Huadong Ma. 2025. Safedriverag: Towards safe autonomous driving with knowledge graph-based retrieval-augmented generation. InProceedings of the 33rd ACM International Conference on Multimedia. 11170–11178

    2025

  38. [46]

    Muchao Ye, Weiyang Liu, and Pan He. 2025. VERA: Explainable Video Anomaly Detection via Verbalized Learning of Vision-Language Models. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

    2025

  39. [47]

    Siyu Yuan, Kaitao Song, Jiangjie Chen, Xu Tan, Dongsheng Li, and Deqing Yang

  40. [48]

    Francesco Zanella et al. 2024. Harnessing Large Language Models for Training- free Video Anomaly Detection. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

    2024

  41. [49]

    Aohan Zeng, Mingdao Liu, Rui Lu, Bowen Wang, Xiao Liu, Yuxiao Dong, and Jie Tang. 2024. Agenttuning: Enabling generalized agent abilities for llms. In Findings of the Association for Computational Linguistics (ACL)

    2024

  42. [50]

    Yirong Zeng et al. 2025. Tool Zero: Training Tool-Augmented LLMs via Pure RL from Scratch. InFindings of the Association for Computational Linguistics (EMNLP)

    2025

  43. [51]

    Huaxin Zhang, Xiaohao Xu, Xiang Wang, Jialong Zuo, Xiaonan Huang, Changxin Gao, Shanjun Zhang, Li Yu, and Nong Sang. 2025. Holmes-vau: Towards long- term video anomaly understanding at any granularity. InProceedings of the computer vision and pattern recognition conference (CVPR)

    2025

  44. [52]

    Yingying Zhang, Desen Zhou, Siqin Chen, Shenghua Gao, and Yi Ma. 2016. Single-image crowd counting via multi-column convolutional neural network. InProceedings of the IEEE conference on computer vision and pattern recognition (CVPR)

    2016

  45. [53]

    Liang Zheng, Liyue Shen, Lu Tian, et al. 2015. Scalable Person Re-Identification: A Benchmark. InProceedings of the IEEE International Conference on Computer Vision (ICCV)

    2015

  46. [54]

    Shu-Qing Zheng, Jian-Huang Zhang, Kai-Qi Huang, et al . 2011. Robust view transformation model for gait recognition. InIEEE International Conference on Image Processing (ICIP)

    2011

  47. [55]

    Yaoyao Zhong, Mengshi Qi, Rui Wang, Yuhan Qiu, Yang Zhang, and Huadong Ma. 2025. Viotgpt: Learning to schedule vision tools towards intelligent video internet of things. InProceedings of the AAAI Conference on Artificial Intelligence, Vol. 39. 10680–10688

    2025

  48. [56]

    Bolei Zhou, Agata Lapedriza, Aditya Khosla, et al . 2017. Places: A 10 Million Image Database for Scene Recognition.IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)40, 6 (2017), 1452–1464

    2017

  49. [57]

    Liyun Zhu, Lei Wang, Arjun Raj, Tom Gedeon, and Chen Chen. 2024. Advancing Video Anomaly Detection: A Concise Review and a New Dataset. InAdvances in Neural Information Processing Systems (NeurIPS)

    2024

  50. [58]

    Xuanyu Zhu, Yuhao Dong, Rundong Wang, Yang Shi, Zhipeng Wu, Yinlun Peng, YiFan Zhang, Yihang Lou, Yuanxing Zhang, Ziwei Liu, et al. 2026. VTC-Bench: Evaluating Agentic Multimodal Models via Compositional Visual Tool Chaining. arXiv preprint arXiv:2603.15030(2026)

    2026

  51. [2024]

    InProceedings of the European Conference on Computer Vision (ECCV)

    Follow the Rules: Reasoning for Video Anomaly Detection with Large Language Models. InProceedings of the European Conference on Computer Vision (ECCV)

  52. [2025]

    InProceedings of the Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL))

    Evoagent: Towards automatic multi-agent generation via evolutionary algorithms. InProceedings of the Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL))

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.