Pith. sign in

REVIEW 3 major objections 4 minor 55 references

FedVAR: Prototype-Aligned Federated Framework for Video Anomaly Recognition

T0 review · 3 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

Pith's one-line read FedVAR claims a single global normality prototype resolves semantic misalignment in federated multi-class video anomaly recognition, outperforming state-of-the-art baselines.

desk verdict A legitimate first federated fine-grained VAR framework with a clean mechanism and honest limitations, but the convergence analysis overreaches in Corollary 1 and the cross-modal subtraction in Eq. 9 is never directly tested. read the letter →

arxiv 2608.06876 v1 pith:T2PP7OV3 submitted 2026-08-07 cs.CV cs.AIcs.DCcs.LG

classification cs.CVcs.AIcs.DCcs.LG MSC 68T4568T0568W15
keywords federatedlearningvideoanomalyrecognitionweaklysupervisedprototypealignmentvision-languagemodelspromptnon-IIDdatasemanticmisalignment
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

FedVAR is a federated framework for fine-grained video anomaly recognition. It identifies semantic misalignment, the divergence of client feature spaces under non-IID data, as the main barrier to multi-class anomaly recognition. Its proposal is to aggregate client-side normality prototypes into one global prototype and re-center every client's visual and textual CLIP features around that shared anchor. If correct, this makes weakly supervised multi-class anomaly recognition feasible in privacy-preserving edge networks without exchanging raw video. The paper reports consistent gains over state-of-the-art federated baselines on UCF-Crime, XD-Violence, and ShanghaiTech across random, event, and scene splits, and it shows transfer to unseen domains and unseen anomaly classes.

What carries the argument

The global normality prototype $m_g = \frac{\sum_i |N_i| m_i}{\sum_i |N_i|}$ is a weighted average of client-side visual prototypes of normal frames. It is broadcast to all clients, and every visual frame feature and every textual anomaly-class prompt is then re-centered by subtracting $m_g$, defining anomaly direction vectors $d_c = E_{\mathrm{text}}([t_{\mathrm{ctx}}, t_c]) - m_g$ that are comparable across clients. Anomaly likelihood is the projection of a re-centered frame onto $d_c$, combined with an Axial Transformer temporal module and a weakly supervised multiple-instance-learning objective.

What would settle it

Take a held-out client and compare, for frames of a given anomaly class c, the alignment between re-centered visual deviations x'_t = E_image(I_t) - m_g and the text direction d_c = E_text([t_ctx, t_c]) - m_g. If the average cosine similarity of true anomalous frames to d_c is not significantly above that of normal frames, or if a zero-shot ranking built from d_c without any training performs at chance, the cross-modal subtraction that the whole alignment rests on is not empirically meaningful.

Watch

Extended reading notes

Core claim

The paper claims to establish that semantic misalignment in federated video anomaly recognition can be resolved by a single shared normality anchor. Each client computes a local prototype from its own normal frames, the server aggregates these into a global prototype, and all clients then re-center both visual frame features and text-prompt embeddings around that anchor. This makes anomaly direction vectors consistent across clients with disjoint anomaly classes and heterogeneous scenes. Empirically, on UCF-Crime, XD-Violence, and ShanghaiTech, under random, event, and scene splits, the method reports state-of-the-art mean AUC and mean AP among federated baselines, and it demonstrates generalization to unseen domains and unseen anomaly classes.

Load-bearing premise

The method assumes that a normal visual prototype computed from normal frames can be subtracted from text embeddings of anomaly class names to produce a meaningful anomaly direction, meaning CLIP's visual and text spaces are additively compatible; if that linearity breaks, the global anchor cannot align clients.

Editorial extensions

If this is right

  • Federated multi-class anomaly recognition becomes achievable with only video-level labels and no raw data exchange between clients.
  • A shared normality anchor provides a transferable representation, yielding improved cross-domain and unseen-class generalization relative to methods that only aggregate prompts or parameters.
  • Communication overhead stays small because the prototype is one D-dimensional vector uploaded once per client, and the trainable module is limited to prompt tokens and a lightweight temporal model.
  • The convergence analysis implies that unaligned local prototypes carry an extra gradient-divergence penalty proportional to prototype spread, so aligned training reaches a target accuracy in no more (and often fewer) communication rounds than the unaligned baseline.
  • The framework stays stable under partial client participation and moderate label noise, since the weighted prototype aggregation filters sparse corruption.

Reading between the lines

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

  • The same prototype re-centering recipe may transfer to other federated tasks with a well-defined "normality" concept, such as industrial defect detection or sensor-based anomaly monitoring, whenever a shared embedding space exists.
  • A direct test of the linearity assumption would be to compare the global-prototype anchor against a text-only prototype or a random vector; the size of the performance gap would reveal how much of the gain comes from the semantic content of the anchor rather than from simple recentering.
  • The paper's passing remark that the prototype can be updated online with an exponential moving average suggests a natural formal extension for handling client churn and long-term distribution shift, which is otherwise left implicit.
  • The slight gap to Fed-WSVAD on XD-Violence suggests that combining prototype alignment with finer temporal localization is a promising extension, a direction the paper itself identifies.
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

3 major / 4 minor

Summary. The paper proposes FedVAR, a federated learning framework for weakly supervised fine-grained video anomaly recognition (VAR). Each client computes a local normality prototype from frozen CLIP image embeddings; the server aggregates these into a global prototype m_g, which is broadcast to all clients and used to re-center both visual frame features and text prompt embeddings before local training of a class-specific prompt context and an axial-temporal transformer. The central claim is that this shared semantic anchor mitigates semantic misalignment under non-IID data and yields state-of-the-art results across UCF-Crime, XD-Violence, and ShanghaiTech under random, event, and scene splits, together with cross-domain and unseen-class generalization and a convergence analysis.

Significance. The problem is well motivated and the proposed mechanism is simple, interpretable, and aligned with the stated privacy constraints. The main empirical support is not circular: Table 9 directly compares global against local prototypes and shows consistent gains under all three partitioning schemes. The evaluation is broad, covering non-IID splits, cross-domain transfer, unseen classes, partial participation, label noise, and communication overhead, and the paper is transparent about several limitations. If the cross-modal cancellation in Eq. (9) is validated, the framework would be a useful new baseline for federated multi-class anomaly recognition. However, the current "consistently outperforms" claim and the comparative convergence guarantee outrun the evidence.

major comments (3)
  1. [Section 4.2, Eq. (9)] The load-bearing assumption that subtracting the visual normality prototype m_g from a text embedding yields a meaningful anomaly direction is never directly tested. Table 9 ablates global versus local prototypes, but both arms subtract the chosen prototype, so the cross-modal cancellation itself is not isolated. Because m_g is computed once and broadcast before training, any systematic error in this text-space subtraction is a shared bias that a frozen text encoder and a single learnable context vector must compensate for, with no guarantee that such compensation exists. I recommend a targeted analysis: compare d_c = E_text([t_ctx, t_c]) - m_g against directions derived from actual anomalous visual features, and include a control ablation that replaces m_g with a random or shifted vector of the same dimension.
  2. [Abstract, Section 7.1, Tables 4, 5, 7, 8] The headline claim that FedVAR "consistently outperforms" state-of-the-art federated baselines is contradicted by the paper's own tables. Fed-WSV AD achieves higher mAP on XD-Violence in Table 4 (49.33 vs. 47.93), higher AP on XD-Violence in Table 5 (77.33 vs. 75.20), higher cross-domain AP in Table 7 (65.42 vs. 60.20), and higher unseen-class AP in Table 8 (77.69 vs. 73.08). Section 8 acknowledges a gap on XD-Violence recognition, but the cross-domain and unseen-class gaps are not acknowledged there. The abstract and conclusion should be revised to state that FedVAR leads on UCF-Crime and ShanghaiTech while remaining competitive on XD-Violence.
  3. [Section 5, Lemma 1 and Corollary 1] The comparative convergence claim is not supported by the stated inequalities. Lemma 1 (Eq. 17) gives an upper bound on the unaligned divergence Gamma^2_local, and Proposition 1 assumes a separate pointwise heterogeneity bound at local prototypes; neither result establishes that the actual Gamma^2_local is larger than Gamma^2_global. Corollary 1 then concludes that the unaligned baseline's error floor is "never smaller" and that more communication rounds are required as Delta_m grows. Comparing upper bounds does not compare actual convergence rates, and a method with a looser bound can perform better. The theory section should be reframed as two independent bounds under different assumptions, and any comparative statement should be derived from matching lower bounds or explicitly added assumptions.
minor comments (4)
  1. [Throughout] The method name is rendered inconsistently as "FedV AR", "FedVAR", and "FedV AR"; please standardize the spelling in the title, abstract, and body.
  2. [Section 7.2, Table 5] The text says the results cover all three partitioning strategies, but Table 5 reports only UCF-Crime and XD-Violence; ShanghaiTech is missing from the table, so the scope should be stated explicitly.
  3. [Section 5, Eq. (20)] Equation (20) uses E both for the number of local epochs and for the expectation operator; adopting a distinct symbol such as E_loc would remove the ambiguity.
  4. [Section 5, Proof Sketches] The proof sketches defer the main induction to references [43, 44]; since Theorem 1 and Proposition 1 are structurally identical to FedAvg with m_g or m_i held fixed, the paper should reproduce the key induction or clearly state which steps are new.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the global-prototype benefit is empirically validated by a direct ablation against local prototypes, and the convergence analysis is a standard FedAvg analysis of the defined objective.

full rationale

FedVAR's derivation chain is self-contained against external benchmarks. The global normality prototype m_g is the sample-size weighted average of client local prototypes m_i (Eqs. 6-7), and both visual and text features are re-centered by subtracting m_g (Eqs. 8-9). The central empirical claim that this shared anchor helps federated VAR is directly tested in Table 9, which compares the global-prototype arm against a local-prototype arm under the same training and evaluation protocol; this is an empirical ablation over held-out test metrics, not a parameter fitted to the target result. The convergence analysis in Section 5 is a standard non-convex FedAvg analysis applied to the explicitly defined objective F(theta; m_g); the paper's own note that 'Theorem 1 below reflects this by construction' is an accurate description of the proof setup, not a hidden reduction of the empirical results to the assumptions. The proof cites standard federated optimization references [43,44], and no load-bearing self-citation is present; the authors' self-citations [30,49] are auxiliary. The cross-modal subtraction assumption in Eq. 9 is an untested modeling assumption that poses a correctness risk, but it is not circular because it is not an input fitted to the reported prediction. No circular step reduces any central claim to its own input, so the score is 0.

Assumptions & free parameters 4 free parameters · 4 assumptions · 1 invented entities

The central method rests on a small set of untested modeling assumptions: cross-modal subtraction in CLIP space, MIL label reliability, and a one-shot global prototype weighted by normal-frame counts. These are not fitted constants but assumptions that, if false, would break the alignment mechanism. The only invented construct, m_g, is data-derived and empirically ablated.

free parameters (4)
  • Per-dataset learning rate eta = 5e-4 (ShanghaiTech), 1e-2 (UCF-Crime, XD-Violence)
    Hand-selected per dataset in Section 6.4. No sensitivity analysis or per-baseline tuning details are reported, so part of the reported margins may reflect hyperparameter choice.
  • FL schedule (rounds R=20, local epochs E=10) = R=20, E=10
    Chosen by hand and used in all main experiments. The convergence bound is asymptotic and does not validate this specific operating point.
  • Video segmentation S=32, F=16 and batch size B=32 = S=32, F=16, B=32
    Manual choices that affect temporal granularity and may matter for sparse anomaly localization.
  • Prompt length t_ctx = 8
    Fixed prompt length with no study of its impact on recognition accuracy.
assumptions (4)
  • domain assumption The CLIP embedding space is linearly composable across modalities: subtracting the visual normality prototype from text embeddings produces valid anomaly direction vectors (Eq. 9).
    This is the foundation of the re-centering step in both AnomalyCLIP and FedVAR. The paper offers no direct validation that visual-to-text subtraction is semantically meaningful.
  • domain assumption Weak video labels satisfy the MIL assumption: an anomalous video contains at least one anomalous frame, and normal-labeled videos contain only normal frames.
    Standard in weakly supervised VAD (Section 3.1), but label quality per client is not verified. Section 7.5.6 only tests swapped frame features, not annotation noise on video-level labels.
  • domain assumption The global prototype m_g, computed once from initial normal-frame counts, remains a valid shared anchor for all clients and all training rounds.
    The paper itself lists this as a limitation in Section 8. Periodic updates are mentioned only as future work, and normal-frame count is used as a quality proxy in Eq. 7.
  • domain assumption Convergence Assumptions 1-4 hold, including the epsilon-separation ||e_c - m|| >= epsilon > 0 used to justify Lipschitzness of g_c(m).
    Assumption 4's justification assumes prompt embeddings stay separated from the prototype; no guarantee is given for trained prompts during optimization.
invented entities (1)
  • Global normality prototype m_g independent evidence
    purpose: A shared semantic anchor obtained by weighted averaging of local normal-frame prototypes; used to re-center visual and textual features at every client.
    Unlike a purely postulated entity, m_g is computed from data and its utility is directly tested by the m_g vs m_i ablation (Table 9) and the PCA illustration (Figure 1). However, its meaning depends on the untested cross-modal subtraction assumption.

how reviews work

0 comments
Cite this review

Pith. "Pith review of FedVAR: Prototype-Aligned Federated Framework for Video Anomaly Recognition." pith.science (2026). https://pith.science/paper/T2PP7OV3

@misc{pith2026260806876,
  author       = {Pith},
  title        = {Pith review of: FedVAR: Prototype-Aligned Federated Framework for Video Anomaly Recognition},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/T2PP7OV3}},
  note         = {Machine review of arXiv:2608.06876}
}
read the original abstract

In the era of Industrial Internet of Things (IIoT) and Cyber-Physical Systems (CPS), Federated Learning (FL) offers a promising decentralized intelligence paradigm for Video Anomaly Recognition (VAR). This task is vital for maintaining high-fidelity Digital Twins and ensuring safety in mission-critical environments. However, the inherent data heterogeneity across distributed edge clients leads to a fundamental challenge known as semantic misalignment, where clients learn divergent feature representations of "normal" and "abnormal" events. The problem becomes particularly pronounced in VAR, where the presence of diverse and fine-grained anomaly categories leads each client to develop distinct semantic interpretations of abnormality. Existing federated methods primarily focus on binary anomaly detection and fail to address this misalignment, preventing effective fine-grained recognition. In this paper, we introduce FedVAR, a weakly-supervised FL framework explicitly designed for VAR. Leveraging the rich representations of Vision-Language Models (VLMs), FedVAR employs a prototype-based alignment mechanism that creates a shared semantic anchor for all clients to re-center and align their visual and textual feature spaces. This process enforces a consistent representation of "normality" across the decentralized network, directly mitigating semantic misalignment and enabling robust prompt-learning of anomaly direction vectors with minimal communication overhead. We conduct extensive experiments on challenging benchmarks under various non-IID data partitioning schemes, unseen domains, and novel anomaly classes. The results demonstrate that FedVAR consistently outperforms state-of-the-art federated baselines, establishing a robust framework for distributed intelligence in video-based CPS.

Figures

Figures reproduced from arXiv: 2608.06876 by the authors.

Figure 1
Figure 1. 3D visualization of normality prototypes, projected into three dimensions using Principal Component Analysis (PCA). Clients learn diverged local [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. The architecture and workflow of our proposed FedVAR framework. Each client computes a local normality prototype ( [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Client-wise data distribution under different partitioning strategies for UCF-Crime and XD-Violence. Each plot reports, for every client, the total number of videos with normal labels and anomalous labels counted separately, where the x-axis denotes client indices and the y-axis indicates the number of videos. Results are shown for three partitioning schemes: random split, event split, and scene split, illustrating … view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Qualitative visualization of anomaly predictions from FedVAR on representative test videos from UCF-Crime, XD-Violence, and ShanghaiTech. For [PITH_FULL_IMAGE:figures/full_fig_p016_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

55 extracted references · 50 canonical work pages

  1. [1]

    Abdalla, S

    M. Abdalla, S. Javed, M. Al Radi, A. Ulhaq, N. Werghi, Video anomaly detection in 10 years: A survey and out- look, Neural Computing and Applications (2025) 1–44

  2. [2]

    Ullah, A

    W. Ullah, A. Ullah, T. Hussain, K. Muhammad, A. A. Heidari, J. Del Ser, S. W. Baik, V . H. C. De Albuquerque, Artificial intelligence of things-assisted two-stream neural network for anomaly detection in surveillance big video data, Future Generation Computer Systems 129 (2022) 286–297

  3. [3]

    Politi, C

    E. Politi, C. Davalas, C. Chronis, G. Dimitrakopoulos, D. Michail, I. Varlamis, Real-time quality monitoring and anomaly detection for vision sensors in connected and au- tonomous vehicles, IEEE Access (2025)

  4. [4]

    Bedogni, M

    L. Bedogni, M. Mamei, M. Picone, M. Pietri, F. Zam- bonelli, Fluid computing & digital twins for intelligent interoperability in the iot ecosystem, Future Generation Computer Systems (2025) 107855. 17

  5. [5]

    Pathirannahalage, V

    I. Pathirannahalage, V . Jayasooriya, J. Samarabandu, A. Subasinghe, A comprehensive analysis of real-time video anomaly detection methods for human and vehicu- lar movement, Multimedia Tools and Applications 84 (10) (2025) 7519–7564

  6. [6]

    Hussain, W

    A. Hussain, W. Ullah, N. Khan, Z. A. Khan, H. Yar, S. W. Baik, Class-incremental learning network for real-time anomaly recognition in surveillance environments, Pattern Recognition 170 (2026) 112064

  7. [7]

    S. Alem, D. Espes, L. Nana, E. Martin, F. De Lamotte, A novel bi-anomaly-based intrusion detection system ap- proach for industry 4.0, Future Generation Computer Sys- tems 145 (2023) 267–283

  8. [8]

    Djebali, G

    S. Djebali, G. Guerard, I. Taleb, Survey and insights on digital twins design and smart grid’s applications, Future Generation Computer Systems 153 (2024) 234–248

Show all 55 references
  1. [9]

    Elshenawy, B

    M. Elshenawy, B. Abdulhai, M. El-Darieby, Towards a service-oriented cyber-physical systems of systems for smart city mobility applications, Future Generation Com- puter Systems 79 (2018) 575–587

  2. [10]

    Sultani, C

    W. Sultani, C. Chen, M. Shah, Real-world anomaly detec- tion in surveillance videos, in: Proceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 6479–6488

  3. [11]

    Karim, K

    H. Karim, K. Doshi, Y . Yilmaz, Real-time weakly su- pervised video anomaly detection, in: Proceedings of the IEEE/CVF winter conference on applications of computer vision, 2024, pp. 6848–6856

  4. [12]

    Radford, J

    A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clark, et al., Learning transferable visual models from natural language supervision, in: International conference on ma- chine learning, PMLR, 2021, pp. 8748–8763

  5. [13]

    Zanella, B

    L. Zanella, B. Liberatori, W. Menapace, F. Poiesi, Y . Wang, E. Ricci, Delving into clip latent space for video anomaly recognition, Computer Vision and Image Under- standing 249 (2024) 104163

  6. [14]

    M. M. Salim, D. Camacho, J. H. Park, Digital twin and federated learning enabled cyberthreat detection system for iot networks, Future Generation Computer Systems 161 (2024) 701–713

  7. [15]

    Barbone, N

    A. Barbone, N. Bicocchi, M. Martinelli, R. Morandi, M. Picone, On-device ai and digital twins: a synergis- tic approach to intelligent cyber-physical systems, Future Generation Computer Systems (2025) 108068

  8. [16]

    McMahan, E

    B. McMahan, E. Moore, D. Ramage, S. Hampson, B. A. y Arcas, Communication-efficient learning of deep net- works from decentralized data, in: Artificial intelligence and statistics, PMLR, 2017, pp. 1273–1282

  9. [17]

    B. Wang, C. Huang, J. Wen, W. Wang, Y . Liu, Y . Xu, Fed- erated weakly supervised video anomaly detection with multimodal prompt, in: Proceedings of the AAAI Confer- ence on Artificial Intelligence, V ol. 39, 2025, pp. 21017– 21025

  10. [18]

    Al-Lahham, M

    A. Al-Lahham, M. Z. Zaheer, N. Tastan, K. Nandaku- mar, Collaborative learning of anomalies with privacy (clap) for unsupervised video anomaly detection: A new baseline, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 12416–12425

  11. [19]

    C. Ju, T. Han, K. Zheng, Y . Zhang, W. Xie, Prompting visual-language models for efficient video understanding, in: European conference on computer vision, Springer, 2022, pp. 105–124

  12. [20]

    Doshi, Y

    K. Doshi, Y . Yilmaz, Privacy-preserving video under- standing via transformer-based federated learning, in: 2023 IEEE Conference on Dependable and Secure Com- puting (DSC), IEEE, 2023, pp. 1–8

  13. [21]

    T. Guo, S. Guo, J. Wang, X. Tang, W. Xu, Promptfl: Let federated participants cooperatively learn prompts instead of models–federated learning in age of foundation model, IEEE Transactions on Mobile Computing 23 (5) (2023) 5179–5194

  14. [22]

    D. Gong, L. Liu, V . Le, B. Saha, M. R. Mansour, S. Venkatesh, A. v. d. Hengel, Memorizing normality to detect anomaly: Memory-augmented deep autoencoder for unsupervised anomaly detection, in: Proceedings of the IEEE/CVF international conference on computer vi- sion, 2019, pp...

  15. [23]

    W. Liu, W. Luo, D. Lian, S. Gao, Future frame prediction for anomaly detection–a new baseline, in: Proceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 6536–6545

  16. [24]

    Qasim, E

    M. Qasim, E. Verdu, Video anomaly detection system us- ing deep convolutional and recurrent models, Results in Engineering 18 (2023) 101026

  17. [25]

    X. Chen, S. Kan, F. Zhang, Y . Cen, L. Zhang, D. Zhang, Multiscale spatial temporal attention graph convolution network for skeleton-based anomaly behavior detection, Journal of visual communication and image representa- tion 90 (2023) 103707

  18. [26]

    Biradar, D

    K. Biradar, D. K. Tyagi, R. B. Battula, P. Subbarao, Ro- bust anomaly detection through transformer-encoded fea- ture diversity learning, in: Proceedings of the Asian Con- ference on Computer Vision, 2024, pp. 115–128

  19. [27]

    P. Wu, X. Zhou, G. Pang, Y . Sun, J. Liu, P. Wang, Y . Zhang, Open-vocabulary video anomaly detection, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 18297–18307. 18

  20. [28]

    H. Lv, Z. Yue, Q. Sun, B. Luo, Z. Cui, H. Zhang, Unbiased multiple instance learning for weakly supervised video anomaly detection, in: Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition, 2023, pp. 8022–8031

  21. [29]

    Z. Lu, H. Pan, Y . Dai, X. Si, Y . Zhang, Federated learn- ing with non-iid data: A survey, IEEE Internet of Things Journal 11 (11) (2024) 19188–19209

  22. [30]

    Kundroo, S

    M. Kundroo, S. H. Kim, T. Kim, Fedchyper: Client- side dynamic hyper-parameter tuning for enhanced feder- ated learning, Journal of Systems Architecture 177 (2026) 103843

  23. [31]

    Xiang, X

    H. Xiang, X. Zhang, X. Xu, A. Beheshti, L. Qi, Y . Hong, W. Dou, Federated learning-based anomaly detection with isolation forest in the iot-edge continuum, ACM Trans- actions on Multimedia Computing, Communications and Applications 22 (1) (2026) 1–19

  24. [32]

    B. Dong, D. Chen, Y . Wu, S. Tang, Y . Zhuang, Fadngs: Federated learning for anomaly detection, IEEE Transac- tions on Neural Networks and Learning Systems (2024)

  25. [33]

    K. Zhou, J. Yang, C. C. Loy, Z. Liu, Learning to prompt for vision-language models, International Journal of Com- puter Vision 130 (9) (2022) 2337–2348

  26. [34]

    K. Zhou, J. Yang, C. C. Loy, Z. Liu, Conditional prompt learning for vision-language models, in: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 16816–16825

  27. [35]

    H. Wang, F. Liu, L. Jiao, J. Wang, Z. Hao, S. Li, L. Li, P. Chen, X. Liu, Vilt-clip: Video and language tuning clip with multimodal prompt learning and scenario-guided op- timization, in: Proceedings of the AAAI Conference on Artificial Intelligence, V ol. 38, 2024, pp. 5390–5400

  28. [36]

    S. Bai, M. Zhang, W. Zhou, S. Huang, Z. Luan, D. Wang, B. Chen, Prompt-based distribution alignment for unsu- pervised domain adaptation, in: Proceedings of the AAAI conference on artificial intelligence, V ol. 38, 2024, pp. 729–737

  29. [37]

    P. Wu, X. Zhou, G. Pang, L. Zhou, Q. Yan, P. Wang, Y . Zhang, Vadclip: Adapting vision-language models for weakly supervised video anomaly detection, in: Proceed- ings of the AAAI Conference on Artificial Intelligence, V ol. 38, 2024, pp. 6074–6082

  30. [38]

    Vaswani, N

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, I. Polosukhin, Attention is all you need, Advances in neural information processing systems 30 (2017)

  31. [39]

    M. Wang, J. Xing, J. Mei, Y . Liu, Y . Jiang, Actionclip: Adapting language-image pretrained models for video ac- tion recognition, IEEE transactions on neural networks and learning systems (2023)

  32. [40]

    T. Guo, S. Guo, J. Wang, Pfedprompt: Learning per- sonalized prompt for vision-language models in federated learning, in: Proceedings of the ACM Web Conference 2023, 2023, pp. 1364–1374

  33. [41]

    Yang, C.-Y

    F.-E. Yang, C.-Y . Wang, Y .-C. F. Wang, Efficient model personalization in federated learning via client-specific prompt generation, in: Proceedings of the IEEE/CVF In- ternational Conference on Computer Vision, 2023, pp. 19159–19168

  34. [42]

    X. Li, M. Jiang, X. Zhang, M. Kamp, Q. Dou, Fedbn: Fed- erated learning on non-iid features via local batch normal- ization, in: International Conference on Learning Repre- sentations (ICLR), 2021

  35. [43]

    S. Wang, T. Tuor, T. Salonidis, K. K. Leung, C. Makaya, T. He, K. Chan, Adaptive federated learning in resource constrained edge computing systems, IEEE journal on se- lected areas in communications 37 (6) (2019) 1205–1221

  36. [44]

    H. Yang, M. Fang, J. Liu, Achieving linear speedup with partial worker participation in non-iid federated learning, in: International Conference on Learning Representations (ICLR), 2021

  37. [45]

    P. Wu, J. Liu, Y . Shi, Y . Sun, F. Shao, Z. Wu, Z. Yang, Not only look, but also listen: Learning multimodal violence detection under weak supervision, in: European confer- ence on computer vision, Springer, 2020, pp. 322–339

  38. [46]

    W. Luo, W. Liu, S. Gao, A revisit of sparse coding based anomaly detection in stacked rnn framework, in: Proceed- ings of the IEEE international conference on computer vi- sion, 2017, pp. 341–349

  39. [47]

    Zhong, N

    J.-X. Zhong, N. Li, W. Kong, S. Liu, T. H. Li, G. Li, Graph convolutional label noise cleaner: Train a plug-and-play action classifier for anomaly detection, in: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2019, pp. 1237–1246

  40. [48]

    Ansel, E

    J. Ansel, E. Yang, H. He, N. Gimelshein, A. Jain, M. V oz- nesensky, B. Bao, P. Bell, D. Berard, E. Burovski, et al., Pytorch 2: Faster machine learning through dynamic python bytecode transformation and graph compilation, in: Proceedings of the 29th ACM International Con- fer...

  41. [49]

    Kundroo, G

    M. Kundroo, G. Haider, N. Khoa, A. W. Mamond, T. Kim, Fedeasy : Federated learning with ease, SoftwareX 31 (2025) 102276

  42. [50]

    J. Ho, N. Kalchbrenner, D. Weissenborn, T. Salimans, Axial attention in multidimensional transformers, arXiv preprint arXiv:1912.12180 (2019). 19

  43. [51]

    Loshchilov, F

    I. Loshchilov, F. Hutter, Decoupled weight decay regular- ization, in: International Conference on Learning Repre- sentations, 2019

  44. [52]

    S. J. Reddi, Z. Charles, M. Zaheer, Z. Garrett, K. Rush, J. Koneˇcn`y, S. Kumar, H. B. McMahan, Adaptive feder- ated optimization, in: International Conference on Learn- ing Representations, 2021

  45. [53]

    D. Yin, Y . Chen, R. Kannan, P. Bartlett, Byzantine-robust distributed learning: Towards optimal statistical rates, in: International Conference On Machine Learning, PMLR, 2018, pp. 5650–5659

  46. [54]

    T.-M. H. Hsu, H. Qi, M. Brown, Measuring the effects of non-identical data distribution for federated visual classi- fication, arXiv preprint arXiv:1909.06335 (2019)

  47. [55]

    Allouah, S

    Y . Allouah, S. Farhadkhani, R. Guerraoui, N. Gupta, R. Pinot, G. Rizk, S. V oitovych, Byzantine-robust feder- ated learning: Impact of client subsampling and local up- dates, in: International Conference on Machine Learning, PMLR, 2024, pp. 1078–1114. 20

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.