Pith. sign in

REVIEW 4 major objections 6 minor 64 references

Token-Level Prompt Mixture with Parameter-Free Routing for Federated Domain Generalization

T0 review · 4 major / 6 minor · reviewed 2026-08-16 · deepseek-v4-flash

Pith's one-line read TRIP claims that routing individual image tokens to specialized prompt experts through a parameter-free optimal-transport assignment gives federated domain generalization that beats single-prompt and image-level methods while transmitting…

desk verdict A solid empirical FedDG paper whose parameter-free token-level router is clever and works, but the semantic-specialization story is overclaimed and the headline communication number is for the lite variant. read the letter →

arxiv 2504.21063 v1 pith:L2HA6KSZ submitted 2025-04-29 cs.LG cs.AI

classification cs.LGcs.AI
keywords federateddomaingeneralizationpromptlearningmixtureofexpertsparameter-freeroutingoptimaltransporttoken-levelCLIPcommunicationefficiency
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

Federated domain generalization trains a shared model across clients that own different visual domains, then expects it to work on a domain none of the clients saw. The paper argues that the usual approach, learning one global text prompt on top of a frozen CLIP model, is too coarse, and that existing prompt mixture-of-experts fixes route whole images and ship learnable routers, which costs communication. TRIP instead routes individual image tokens to prompt experts: tokens are first clustered with a capacity constraint, clusters are matched to static orthogonal keys by optimal transport, and each image's prompt is a token-count-weighted blend of the experts. A KL term pulls each client's predictions toward CLIP's zero-shot distribution to stop prompts from overfitting their local domain. If the reported numbers hold, prompt-based federated domain generalization gets state-of-the-art accuracy on PACS, Office-Home, VLCS, and DomainNet while transmitting only prompt parameters, as few as 1K per round in the lite configuration.

What carries the argument

The load-bearing object is the parameter-free router. It has three parts: capacity-aware $k$-means clustering over CLIP's token embeddings, which groups patches into balanced clusters constrained by a capacity factor $\alpha$; a set of $M$ static, randomly initialized, mutually orthogonal key vectors, each permanently associated with one prompt expert; and an optimal-transport assignment, implemented with the Hungarian algorithm, that matches cluster centroids to keys using a cosine-distance cost matrix. Because the keys are fixed while the prompts evolve, the routing stays stable across training iterations and across clients, and because the router has no trainable parameters, nothing but the expert prompts needs to be communicated.

What would settle it

Re-run TRIP on PACS or Office-Home with the static keys re-randomized every communication round, keeping all other settings fixed; a large accuracy drop would confirm that stable key anchoring carries the result, while no drop would indicate the routing mechanism itself is not the source of the gain.

Watch

Extended reading notes

Core claim

The central claim is that token-level, parameter-free routing is both feasible and sufficient for state-of-the-art federated domain generalization with vision-language models. Treating each of $M$ learnable prompts as an expert, TRIP routes the patch tokens of every image to experts rather than routing the whole image, then forms the image's prompt as $\sum_i \pi_i E_i$ with $\pi_i$ the fraction of tokens sent to expert $i$. The routing is parameter-free: capacity-aware $k$-means groups tokens into balanced clusters, and optimal transport, solved by the Hungarian algorithm, assigns clusters to fixed orthogonal keys, one key per expert, so that semantically similar clusters land on the same expert even as the prompt experts themselves change during training. The unbiased learning objective trains experts with cross-entropy plus a KL-divergence penalty that anchors local predictions to CLIP's zero-shot distribution. On the paper's evidence, this combination outperforms single-global-prompt methods, image-level mixture-of-experts methods, and learnable routers on four benchmarks while communicating only the prompt parameters.

Load-bearing premise

The load-bearing premise is that k-means over CLIP's frozen patch-token embeddings groups patches into the same semantic specialities the experts should learn, so the static keys, which are never trained, remain sensible anchors.

Editorial extensions

If this is right

  • TRIP reports higher average accuracy than the compared prompt-learning federated methods on all four benchmarks, with gains over the strongest baseline between 0.38 and 2.02 percentage points depending on the dataset.
  • Token-level routing outperforms image-level routing in the paper's comparison, for both learnable and parameter-free routers, which isolates fine-grained token assignment as the source of improvement.
  • Removing the KL debiasing term hurts accuracy on PACS, Office-Home, and VLCS, so anchoring local prompt experts to the vision-language model's zero-shot distribution is a necessary part of the method.
  • Static orthogonal keys keep the clustering-based router stable across random seeds, so the routing does not add noticeable inference variance despite being re-computed from scratch.
  • TRIP-Lite communicates about 1,000 parameters per round and still matches or exceeds methods that transmit thousands to millions of parameters, so the communication bottleneck is effectively removed.

Reading between the lines

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

  • Inference: because the router's keys never depend on the trained experts, the same clustering-plus-optimal-transport assignment could route any set of parameter-efficient modules, such as adapters or LoRA, not just text prompts.
  • Inference: the paper fixes the inference capacity factor at 2.0 for every target domain; tuning it per target domain, or making it depend on how many tokens an image produces, is a natural extension the paper leaves open.
  • Inference: the KL anchor only works when the label space is shared and CLIP's zero-shot distribution is meaningful, so the paper's stated open-set limitation follows directly, since open-set targets would need a different global reference distribution.
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 / 6 minor

Summary. The paper proposes TRIP, a federated domain generalization (FedDG) method built on CLIP prompt learning. TRIP treats multiple prompt vectors as experts and performs token-level routing: image tokens are grouped by a capacity-constrained k-means clustering, clusters are matched to prompt experts via an optimal-transport-style assignment that uses fixed random orthogonal keys rather than trainable router parameters, and the instance-specific prompt is the token-count-weighted average of the experts. A KL regularizer aligns local prediction distributions with CLIP's zero-shot distribution to reduce client-specific bias. Experiments on PACS, Office-Home, VLCS, and DomainNet report improved accuracy over CL-based, FL-based, conventional FedDG, and prompt-based baselines, with low communication cost; the lightweight TRIP-Lite configuration communicates about 1K parameters per round. Ablations, hyperparameter studies, stability checks, and Grad-CAM visualizations support the main design choices.

Significance. If the empirical results hold, TRIP is a meaningful contribution: it is a token-level prompt mixture for FedDG whose router adds no trainable parameters, and the reported gains over strong prompt-based baselines are consistent across four benchmarks. The paper also provides ablations, a stability analysis, a fair-communication-budget comparison, and public code, which strengthen reproducibility. These strengths make the work a useful step for communication-efficient federated prompt learning. However, the significance is tempered by the overstatement of the '1K parameters per round' claim for the standard configuration, the absence of variance estimates for the headline numbers, and the fact that the semantic-specialization interpretation of the routing is not directly tested.

major comments (4)
  1. [Abstract and Section V-D, Tables I-IV] The abstract and the contribution bullet state that TRIP achieves state-of-the-art results with 'communication of only 1K parameters per round,' but this number applies only to TRIP-Lite (2 experts x 1 token x 512 = 1,024 parameters). The standard TRIP configuration (4 experts x 32 tokens x 512) transmits 65,536 parameters, i.e., 0.065M, per round. The paper should explicitly attribute the 1K claim to TRIP-Lite and report the per-configuration communication costs.
  2. [Section IV-B, Eq. (7), and Section V-E, Table V] The claimed semantic specialization of the routing is not tested against a non-semantic control. The static keys are randomly initialized and never updated, and the cost matrix is the cosine distance to these keys. This design guarantees stable assignment of similar clusters to the same key, but stability is not the same as semantic meaningfulness. The ablations remove clustering entirely or replace keys with dynamic experts, but they never keep clustering fixed and compare the random-key assignment with semantically initialized keys (e.g., CLIP text embeddings of class names) or with a fixed random permutation of cluster-to-expert assignment. Without such a control, the statement in Section IV-B that 'semantically similar clusters... are assigned to the same keys' and the expert-specialization interpretation of Fig. 8 remain unsupported. Please add this control or soften the corresponding claims.
  3. [Section V-D, Tables I-IV] No variance information is reported for the main results. The text says all results are averages of three runs, but standard deviations are absent. Several reported margins are very small (e.g., PACS TRIP-Lite vs. FedAPT: 96.58 vs. 96.52; VLCS: 83.41 vs. 83.22; DomainNet: 58.62 vs. 58.14), so without error bars or significance tests the state-of-the-art claim is not fully established. Please report standard deviations or confidence intervals, or provide per-run results.
  4. [Section V-B, V-C, and Tables I-IV] The baseline comparison needs clarification. The CL-based DG methods are cited from [57] rather than reimplemented; please state explicitly whether they use the same CLIP ViT-B/16 backbone and the same evaluation protocol. More importantly, the reported communication cost of FedAPT changes across tables (0.059M in Table I, 0.534M in Table II, 0.043M in Table III, 2.829M in Table IV). Please explain the source of this variation or correct the numbers, since communication cost is a central comparison.
minor comments (6)
  1. [Section IV-B, Eq. (8)] The notation in Eq. (8) is confusing: the index i is used both for the input image and for the expert index in the sum. Please use different indices, e.g., E_i^k = sum_{m=1}^M pi_m E_m^k.
  2. [Section IV-B] The method described by Eqs. (5)-(6) with binary Gamma and Hungarian algorithm is a linear assignment problem, not optimal transport in the usual relaxed sense. Clarifying this terminology would avoid confusion.
  3. [Section V-H and Fig. 4] The inference-stability box plots are shown only for Office-Home and VLCS. Please either include PACS and DomainNet or state why they are omitted.
  4. [Fig. 3] The bar chart comparing image-level and token-level routers does not show error bars or exact numerical values. Adding them would make the comparison more informative.
  5. [Table V] The ablation table removes components cumulatively, but the caption does not say so. Please make explicit that each row drops the components marked with x relative to the full model.
  6. [Algorithm 1] Step 2 sends the keys to all clients only once. Please clarify whether the keys are also needed at inference time and note that this one-time transmission is not counted in the per-round communication cost.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: routing and debiasing rely on fixed references, and the self-citation is not load-bearing.

full rationale

TRIP's derivation chain is self-contained. The routing cost in Eq. (7) is computed from frozen CLIP token centroids and randomly initialized, never-updated orthogonal keys; it is not optimized to reproduce labels or the reported accuracies. The instance-specific prompt in Eq. (8) is a weighted average of prompt experts whose weights come from token counts, not from any fitted target. The debiasing loss in Eq. (11) anchors local predictions to CLIP's zero-shot distribution, which is a fixed external reference, rather than to unseen target results. The FedAvg aggregation in Eq. (13) is standard. The paper cites its own prior prompt-learning work [11] for the FedDG prompt-learning context and for dataset split ratios, but neither use is load-bearing: disbelieving [11] would not change the routing mechanism, the KL regularizer, or the benchmark protocol, which is also anchored to external sources [53], [54]. There is no imported uniqueness theorem, no ansatz smuggled in via self-citation, and no fitted parameter renamed as a prediction. The skeptical concern that random orthogonal keys may make the Hungarian assignment nearly arbitrary is a substantive correctness risk about whether the routing is semantically meaningful, but it is not circularity: the routing is not defined in terms of the outcomes it is used to predict.

Assumptions & free parameters 5 free parameters · 3 assumptions · 1 invented entities

The central empirical claims rest on several tuned hyperparameters (alpha, beta, expert count, token count, key initialization) and on the assumption that CLIP's frozen representations and zero-shot distribution are reliable. No new physical or mathematical entity is postulated beyond the static routing keys.

free parameters (5)
  • capacity factor alpha = 1.0 (training), 2.0 (inference)
    Controls cluster capacity; inference value selected after observing validation curve in Fig. 5.
  • KL loss weight beta = 0.8
    Balances cross-entropy and KL regularizer; tuned via Table VII.
  • number of prompt experts M = 4 (TRIP), 2 (TRIP-Lite)
    Chosen via ablation in Fig. 6b.
  • number of prompt tokens per expert = 32 (TRIP), 1 (TRIP-Lite)
    Chosen via ablation in Fig. 6a.
  • static key initialization strategy = orthogonal
    Selected via comparison in Table VI.
assumptions (3)
  • domain assumption CLIP zero-shot prediction distribution is a reliable global anchor for reducing local bias
    Used in Section IV-C to define the KL regularizer; if CLIP's zero-shot is unreliable on a target domain, the regularizer could hurt.
  • domain assumption Source and target clients share the same label space
    Stated explicitly in the Conclusion as a limitation of the method.
  • standard math Standard k-means, optimal transport, and Hungarian algorithm behave as described
    The routing mechanism relies on these classical tools without new guarantees.
invented entities (1)
  • Static non-learnable keys
    purpose: Stable anchor vectors that map token clusters to prompt experts via optimal transport, avoiding routing instability from changing experts
    These are algorithmic constructs introduced by the paper. There is no external falsifiable handle; their effectiveness is measured only through downstream accuracy.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Token-Level Prompt Mixture with Parameter-Free Routing for Federated Domain Generalization." pith.science (2026). https://pith.science/paper/L2HA6KSZ

@misc{pith2026250421063,
  author       = {Pith},
  title        = {Pith review of: Token-Level Prompt Mixture with Parameter-Free Routing for Federated Domain Generalization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/L2HA6KSZ}},
  note         = {Machine review of arXiv:2504.21063}
}
read the original abstract

Federated domain generalization (FedDG) aims to learn a globally generalizable model from decentralized clients with heterogeneous data while preserving privacy. Recent studies have introduced prompt learning to adapt vision-language models (VLMs) in FedDG by learning a single global prompt. However, such a one-prompt-fits-all learning paradigm typically leads to performance degradation on personalized samples. Although the mixture of experts (MoE) offers a promising solution for specialization, existing MoE-based methods suffer from coarse image-level expert assignment and high communication costs from parameterized routers. To address these limitations, we propose TRIP, a Token-level prompt mixture with parameter-free routing framework for FedDG, which treats multiple prompts as distinct experts. Unlike existing image-level routing designs, TRIP assigns different tokens within an image to specific experts. To ensure communication efficiency, TRIP incorporates a parameter-free routing mechanism based on token clustering and optimal transport. The instance-specific prompt is then synthesized by aggregating experts, weighted by the number of tokens assigned to each. Additionally, TRIP develops an unbiased learning strategy for prompt experts, leveraging the VLM's zero-shot generalization capability. Extensive experiments across four benchmarks demonstrate that TRIP achieves optimal generalization results, with communication of only 1K parameters per round. Our code is available at https://github.com/GongShuai8210/TRIP.

Figures

Figures reproduced from arXiv: 2504.21063 by the authors.

Figure 1
Figure 1. Comparison of our method, TRIP, with prior prompt-based methods. (a) PromptFL [ [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Conceptual framework of TRIP for FedDG. TRIP consists of two essential components: (a) a parameter-free routing mechanism and (b) unbiased [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Comparison of learnable (L) and our parameter-free (P) image-level [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (5 more)
Figure 5
Figure 5. Figure 5: Effect of capacity factor α during training and inference. TABLE VII EFFECT OF THE KL LOSS COEFFICIENT β. β Art Clipart Product Real Avg. 0.02 83.93 71.75 92.77 90.91 84.84 0.2 84.46 71.91 92.72 91.46 85.14 0.4 84.87 71.01 92.85 91.75 85.12 0.6 85.00 71.13 92.88 91.78 …
Figure 4
Figure 4. Figure 4: Visualization of model inference stability under different random seeds [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 6
Figure 6. Figure 6: Effect of the number of prompt tokens and experts. [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]
Figure 7
Figure 7. Figure 7: Accuracy of PromptFL and TRIP under equivalent communication [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 8
Figure 8. Figure 8: The Grad-CAM visualization results for different prompting methods evaluated on DomainNet. [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

64 extracted references · 17 canonical work pages

  1. [57]

    Hcvp: Leveraging hierarchical contrastive visual prompt for domain generalization,

    G. Zhou, Z. Han, S. Chen, B. Huang, L. Zhu, T. Liu, L. Yao, and K. Zhang, “Hcvp: Leveraging hierarchical contrastive visual prompt for domain generalization,” IEEE Transactions on Multimedia , 2024

  2. [1]

    Communication-efficient learning of deep networks from decentralized data,

    B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. y Arcas, “Communication-efficient learning of deep networks from decentralized data,” in Artificial intelligence and statistics . PMLR, 2017, pp. 1273– 1282. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 12 Cat CLIP TRIP Expert #1 Expert #2 Bicycle CLIP TRIP Expert #1 Expert #2 Flower CLIP...

  3. [2]

    A survey on federated learning systems: Vision, hype and reality for data privacy and protection,

    Q. Li, Z. Wen, Z. Wu, S. Hu, N. Wang, Y . Li, X. Liu, and B. He, “A survey on federated learning systems: Vision, hype and reality for data privacy and protection,” IEEE Transactions on Knowledge and Data Engineering, vol. 35, no. 4, pp. 3347–3366, 2021

  4. [3]

    Mixstyle neural networks for domain generalization and adaptation,

    K. Zhou, Y . Yang, Y . Qiao, and T. Xiang, “Mixstyle neural networks for domain generalization and adaptation,” International Journal of Computer Vision, vol. 132, no. 3, pp. 822–836, 2024

  5. [4]

    Madg: Margin-based adversarial learning for domain generalization,

    A. Dayal, V . KB, L. R. Cenkeramaddi, C. Mohan, A. Kumar, and V . N Balasubramanian, “Madg: Margin-based adversarial learning for domain generalization,” Advances in Neural Information Processing Systems, vol. 36, 2024

  6. [5]

    Feddg: Federated do- main generalization on medical image segmentation via episodic learn- ing in continuous frequency space,

    Q. Liu, C. Chen, J. Qin, Q. Dou, and P.-A. Heng, “Feddg: Federated do- main generalization on medical image segmentation via episodic learn- ing in continuous frequency space,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2021, pp. 1013–1023

  7. [6]

    Learning transferable visual models from natural language supervision,

    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 machine learning . PMLR, 2021, pp. 8748–8763

  8. [7]

    Blip: Bootstrapping language-image pre-training for unified vision-language understanding and generation,

    J. Li, D. Li, C. Xiong, and S. Hoi, “Blip: Bootstrapping language-image pre-training for unified vision-language understanding and generation,” in International conference on machine learning . PMLR, 2022, pp. 12 888–12 900

Show all 64 references
  1. [8]

    Learning to prompt for vision- language models,

    K. Zhou, J. Yang, C. C. Loy, and Z. Liu, “Learning to prompt for vision- language models,” International Journal of Computer Vision , vol. 130, no. 9, pp. 2337–2348, 2022

  2. [9]

    Maple: Multi-modal prompt learning,

    M. U. Khattak, H. Rasheed, M. Maaz, S. Khan, and F. S. Khan, “Maple: Multi-modal prompt learning,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2023, pp. 19 113–19 122

  3. [10]

    Diprompt: Disentangled prompt tuning for multiple latent domain generalization in federated learning,

    S. Bai, J. Zhang, S. Li, S. Guo, J. Guo, J. Hou, T. Han, and X. Lu, “Diprompt: Disentangled prompt tuning for multiple latent domain generalization in federated learning,” arXiv preprint arXiv:2403.08506 , 2024

  4. [11]

    Federated domain generalization via prompt learning and aggregation,

    S. Gong, C. Cui, C. Zhang, W. Wang, X. Nie, and L. Zhu, “Federated domain generalization via prompt learning and aggregation,” arXiv preprint arXiv:2411.10063, 2024

  5. [12]

    Promptfl: Let federated participants cooperatively learn prompts instead of models-federated learning in age of foundation model,

    T. Guo, S. Guo, J. Wang, X. Tang, and W. Xu, “Promptfl: Let federated participants cooperatively learn prompts instead of models-federated learning in age of foundation model,” IEEE Transactions on Mobile Computing, 2023

  6. [13]

    Mixture of experts made personalized: Federated prompt learning for vision-language models,

    J. Luo, C. Chen, and S. Wu, “Mixture of experts made personalized: Federated prompt learning for vision-language models,” arXiv preprint arXiv:2410.10114, 2024

  7. [14]

    Towards instance-adaptive inference for federated learning,

    C.-M. Feng, K. Yu, N. Liu, X. Xu, S. Khan, and W. Zuo, “Towards instance-adaptive inference for federated learning,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2023, pp. 23 287–23 296

  8. [15]

    Adaptive mixtures of local experts,

    R. A. Jacobs, M. I. Jordan, S. J. Nowlan, and G. E. Hinton, “Adaptive mixtures of local experts,” Neural computation, vol. 3, no. 1, pp. 79–87, 1991

  9. [16]

    Deepseekmoe: Towards ultimate expert special- ization in mixture-of-experts language models,

    D. Dai, C. Deng, C. Zhao, R. Xu, H. Gao, D. Chen, J. Li, W. Zeng, X. Yu, Y . Wu et al. , “Deepseekmoe: Towards ultimate expert special- ization in mixture-of-experts language models,” CoRR, 2024

  10. [17]

    Mixture-of-prompt-experts for multi-modal semantic understanding,

    Z. Wu, H.-Y . Huang, F. Qu, and Y . Wu, “Mixture-of-prompt-experts for multi-modal semantic understanding,” in Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024) , 2024, pp. 11 381– 11 393

  11. [18]

    M ´emoire sur la th ´eorie des d ´eblais et des remblais,

    G. Monge, “M ´emoire sur la th ´eorie des d ´eblais et des remblais,” Mem. Math. Phys. Acad. Royale Sci. , pp. 666–704, 1781

  12. [19]

    Self-regulating prompts: Foundational model adaptation without forgetting,

    M. U. Khattak, S. T. Wasim, M. Naseer, S. Khan, M.-H. Yang, and F. S. Khan, “Self-regulating prompts: Foundational model adaptation without forgetting,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2023, pp. 15 190–15 200

  13. [20]

    Federated optimization in heterogeneous networks,

    T. Li, A. K. Sahu, M. Zaheer, M. Sanjabi, A. Talwalkar, and V . Smith, “Federated optimization in heterogeneous networks,” Proceedings of Machine learning and systems , vol. 2, pp. 429–450, 2020

  14. [21]

    Fedfame: A data augmentation free framework based on model contrastive learning for federated semi-supervised learning,

    S. Malaviya, M. Shukla, P. Korat, and S. Lodha, “Fedfame: A data augmentation free framework based on model contrastive learning for federated semi-supervised learning,” in Proceedings of the 38th ACM/SI- GAPP Symposium on Applied Computing , 2023, pp. 1114–1121

  15. [22]

    Scaffold: Stochastic controlled averaging for federated learn- ing,

    S. P. Karimireddy, S. Kale, M. Mohri, S. Reddi, S. Stich, and A. T. Suresh, “Scaffold: Stochastic controlled averaging for federated learn- ing,” in International conference on machine learning . PMLR, 2020, pp. 5132–5143

  16. [23]

    Federated learning with matched averaging,

    H. Wang, M. Yurochkin, Y . Sun, D. Papailiopoulos, and Y . Khazaeni, “Federated learning with matched averaging,” in International Confer- ence on Learning Representations , 2020

  17. [24]

    Tackling the objective inconsistency problem in heterogeneous federated optimiza- tion,

    J. Wang, Q. Liu, H. Liang, G. Joshi, and H. V . Poor, “Tackling the objective inconsistency problem in heterogeneous federated optimiza- tion,” Advances in neural information processing systems , vol. 33, pp. 7611–7623, 2020

  18. [25]

    Fedbn: Feder- ated learning on non-iid features via local batch normalization,

    X. Li, M. Jiang, X. Zhang, M. Kamp, and Q. Dou, “Fedbn: Feder- ated learning on non-iid features via local batch normalization,” arXiv preprint arXiv:2102.07623, 2021

  19. [26]

    Fair federated learning under domain skew with local consistency and domain diversity,

    Y . Chen, W. Huang, and M. Ye, “Fair federated learning under domain skew with local consistency and domain diversity,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 12 077–12 086

  20. [27]

    Domain generalization with adversarial feature learning,

    H. Li, S. J. Pan, S. Wang, and A. C. Kot, “Domain generalization with adversarial feature learning,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2018, pp. 5400–5409

  21. [28]

    Unified deep supervised domain adaptation and generalization,

    S. Motiian, M. Piccirilli, D. A. Adjeroh, and G. Doretto, “Unified deep supervised domain adaptation and generalization,” in Proceedings of the IEEE international conference on computer vision , 2017, pp. 5715– 5725

  22. [29]

    Domain generalization with mixstyle,

    K. Zhou, Y . Yang, Y . Qiao, and T. Xiang, “Domain generalization with mixstyle,” in International Conference on Learning Representations

  23. [30]

    Learning to optimize domain specific normalization for domain generalization,

    S. Seo, Y . Suh, D. Kim, G. Kim, J. Han, and B. Han, “Learning to optimize domain specific normalization for domain generalization,” in Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXII 16 . Springer, 2020, pp. 68–83

  24. [31]

    Learning to generalize: Meta-learning for domain generalization,

    D. Li, Y . Yang, Y .-Z. Song, and T. Hospedales, “Learning to generalize: Meta-learning for domain generalization,” in Proceedings of the AAAI conference on artificial intelligence , vol. 32, no. 1, 2018

  25. [32]

    Do- main generalization via model-agnostic learning of semantic features,

    Q. Dou, D. Coelho de Castro, K. Kamnitsas, and B. Glocker, “Do- main generalization via model-agnostic learning of semantic features,” Advances in neural information processing systems , vol. 32, 2019

  26. [33]

    Federated domain generaliza- tion for image recognition via cross-client style transfer,

    J. Chen, M. Jiang, Q. Dou, and Q. Chen, “Federated domain generaliza- tion for image recognition via cross-client style transfer,” in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2023, pp. 361–370

  27. [34]

    Stablefdg: style and attention based learning for federated domain generalization,

    J. Park, D.-J. Han, J. Kim, S. Wang, C. Brinton, and J. Moon, “Stablefdg: style and attention based learning for federated domain generalization,” Advances in Neural Information Processing Systems , vol. 36, 2024

  28. [35]

    Rethinking federated learning with domain shift: A prototype view,

    W. Huang, M. Ye, Z. Shi, H. Li, and B. Du, “Rethinking federated learning with domain shift: A prototype view,” in 2023 IEEE/CVF Con- JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 13 ference on Computer Vision and Pattern Recognition (CVPR) . IEEE, 2023, pp. 16 312–16 322

  29. [36]

    Federated domain generalization with generalization adjustment,

    R. Zhang, Q. Xu, J. Yao, Y . Zhang, Q. Tian, and Y . Wang, “Federated domain generalization with generalization adjustment,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion, 2023, pp. 3954–3963

  30. [37]

    Conditional prompt learning for vision-language models,

    K. Zhou, J. Yang, C. C. Loy, and Z. Liu, “Conditional prompt learning for vision-language models,” in Proceedings of the IEEE/CVF confer- ence on computer vision and pattern recognition , 2022, pp. 16 816– 16 825

  31. [38]

    Plot: Prompt learning with optimal transport for vision-language models,

    G. Chen, W. Yao, X. Song, X. Li, Y . Rao, and K. Zhang, “Plot: Prompt learning with optimal transport for vision-language models,” in The Eleventh International Conference on Learning Representations

  32. [39]

    Tuning multi-mode token-level prompt alignment across modalities,

    D. Wang, M. Li, X. Liu, M. Xu, B. Chen, and H. Zhang, “Tuning multi-mode token-level prompt alignment across modalities,” Advances in Neural Information Processing Systems , vol. 36, 2024

  33. [40]

    Global and local prompts coop- eration via optimal transport for federated learning,

    H. Li, W. Huang, J. Wang, and Y . Shi, “Global and local prompts coop- eration via optimal transport for federated learning,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 12 151–12 161

  34. [41]

    Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity,

    W. Fedus, B. Zoph, and N. Shazeer, “Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity,” Journal of Machine Learning Research , vol. 23, no. 120, pp. 1–39, 2022

  35. [42]

    Mixture of lora experts,

    X. Wu, S. Huang, and F. Wei, “Mixture of lora experts,” in The Twelfth International Conference on Learning Representations

  36. [43]

    Nonlinear models using dirichlet process mixtures

    B. Shahbaba and R. Neal, “Nonlinear models using dirichlet process mixtures.” Journal of Machine Learning Research , vol. 10, no. 8, 2009

  37. [44]

    Outrageously large neural networks: The sparsely-gated mixture-of-experts layer,

    N. Shazeer, A. Mirhoseini, K. Maziarz, A. Davis, Q. Le, G. Hinton, and J. Dean, “Outrageously large neural networks: The sparsely-gated mixture-of-experts layer,” arXiv preprint arXiv:1701.06538 , 2017

  38. [45]

    Fedjets: Efficient just-in-time personalization with federated mixture of experts,

    C. Dun, M. H. Garcia, G. Zheng, A. Awadallah, R. Sim, A. Kyrillidis, and D. Dimitriadis, “Fedjets: Efficient just-in-time personalization with federated mixture of experts,” in R0-FoMo: Robustness of Few-shot and Zero-shot Learning in Large Foundation Models , 2023

  39. [46]

    Federated mixture of experts,

    M. Reisser, C. Louizos, E. Gavves, and M. Welling, “Federated mixture of experts,” arXiv preprint arXiv:2107.06724 , 2021

  40. [47]

    Language models are unsupervised multitask learners,

    A. Radford, J. Wu, R. Child, D. Luan, D. Amodei, I. Sutskever et al., “Language models are unsupervised multitask learners,” OpenAI blog , vol. 1, no. 8, p. 9, 2019

  41. [48]

    The hungarian method for the assignment problem,

    H. W. Kuhn, “The hungarian method for the assignment problem,” Naval research logistics quarterly, vol. 2, no. 1-2, pp. 83–97, 1955

  42. [49]

    Deeper, broader and artier domain generalization,

    D. Li, Y . Yang, Y .-Z. Song, and T. M. Hospedales, “Deeper, broader and artier domain generalization,” in Proceedings of the IEEE international conference on computer vision , 2017, pp. 5542–5550

  43. [50]

    Deep hashing network for unsupervised domain adaptation,

    H. Venkateswara, J. Eusebio, S. Chakraborty, and S. Panchanathan, “Deep hashing network for unsupervised domain adaptation,” in Pro- ceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 5018–5027

  44. [51]

    Unbiased metric learning: On the utilization of multiple datasets and web images for softening bias,

    C. Fang, Y . Xu, and D. N. Rockmore, “Unbiased metric learning: On the utilization of multiple datasets and web images for softening bias,” in Proceedings of the IEEE International Conference on Computer Vision , 2013, pp. 1657–1664

  45. [52]

    Moment matching for multi-source domain adaptation,

    X. Peng, Q. Bai, X. Xia, Z. Huang, K. Saenko, and B. Wang, “Moment matching for multi-source domain adaptation,” in Proceedings of the IEEE/CVF international conference on computer vision, 2019, pp. 1406– 1415

  46. [53]

    Self-challenging im- proves cross-domain generalization,

    Z. Huang, H. Wang, E. P. Xing, and D. Huang, “Self-challenging im- proves cross-domain generalization,” in Computer vision–ECCV 2020: 16th European conference, Glasgow, UK, August 23–28, 2020, proceed- ings, part II 16 . Springer, 2020, pp. 124–140

  47. [54]

    A fourier-based framework for domain generalization,

    Q. Xu, R. Zhang, Y . Zhang, Y . Wang, and Q. Tian, “A fourier-based framework for domain generalization,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2021, pp. 14 383–14 392

  48. [55]

    Swad: Domain generalization by seeking flat minima,

    J. Cha, S. Chun, K. Lee, H.-C. Cho, S. Park, Y . Lee, and S. Park, “Swad: Domain generalization by seeking flat minima,” Advances in Neural Information Processing Systems , vol. 34, pp. 22 405–22 418, 2021

  49. [56]

    Prompt vision transformer for domain generalization,

    Z. Zheng, X. Yue, K. Wang, and Y . You, “Prompt vision transformer for domain generalization,” arXiv preprint arXiv:2208.08914 , 2022

  50. [58]

    Fedsr: A simple and effective domain generalization method for federated learning,

    A. T. Nguyen, P. Torr, and S. N. Lim, “Fedsr: A simple and effective domain generalization method for federated learning,” Advances in Neural Information Processing Systems , vol. 35, pp. 38 831–38 843, 2022

  51. [59]

    Federated learning with domain generalization,

    L. Zhang, X. Lei, Y . Shi, H. Huang, and C. Chen, “Federated learning with domain generalization,” arXiv preprint arXiv:2111.10487 , 2021

  52. [60]

    Fedclip: Fast generalization and personalization for clip in federated learning,

    W. Lu, H. Xixu, J. Wang, and X. Xie, “Fedclip: Fast generalization and personalization for clip in federated learning,” in ICLR 2023 Workshop on Trustworthy and Reliable Large-Scale Machine Learning Models , 2023

  53. [61]

    Federated adaptive prompt tuning for multi-domain collaborative learning,

    S. Su, M. Yang, B. Li, and X. Xue, “Federated adaptive prompt tuning for multi-domain collaborative learning,” in Proceedings of the AAAI Conference on Artificial Intelligence , vol. 38, no. 13, 2024, pp. 15 117– 15 125

  54. [62]

    An image is worth 16x16 words: Transformers for image recognition at scale,

    A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Minderer, G. Heigold, S. Gelly et al., “An image is worth 16x16 words: Transformers for image recognition at scale,” in International Conference on Learning Representations , 2020

  55. [63]

    Grad-cam: Visual explanations from deep networks via gradient-based localization,

    R. R. Selvaraju, M. Cogswell, A. Das, R. Vedantam, D. Parikh, and D. Batra, “Grad-cam: Visual explanations from deep networks via gradient-based localization,” in Proceedings of the IEEE international conference on computer vision , 2017, pp. 618–626. Shuai Gong received his m...

  56. [2021]

    After graduation, he was worked as a Postdoc Research Associate in School of Computer Science, Carnegie Mellon University, a Senior Lecturer in Faculty of Information Technology, Monash University, and an Associate Professor in School of Computing Technologies, RMIT University...

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.