Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

Componential Prompt-Knowledge Alignment for Domain Incremental Learning

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read The paper claims that component-wise misalignment between domain-specific prompts causes conflicting knowledge fusion, and that KA-Prompt's mined initialization plus online alignment loss fixes it, gaining 4.25–5.59 accuracy points over…

desk verdict A solid empirical prompt-based DIL paper with consistent gains over C-Prompt, but the component-alignment story is more asserted than proven; the auxiliary loss is a regularizer that likely helps for other reasons. read the letter →

arxiv 2505.04575 v1 pith:RMIKU23K submitted 2025-05-07 cs.CV cs.LG

classification cs.CVcs.LG
keywords domainincrementallearningprompt-basedmethodscomponent-wisepromptalignmentreusableknowledgemininggreedysearchonlinevisiontransformer
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

Domain-incremental learning systems that store knowledge in per-domain prompts risk mixing unrelated knowledge when prompts are averaged at inference time, because the components inside each prompt are learned independently and end up in different positions across domains. The paper argues this component-wise misalignment is an intrinsic limit of methods like C-Prompt, and that it also prevents new domains from reusing generalizable knowledge stored in old prompts. KA-Prompt addresses both problems: a greedy search mines a small set of old prompts whose knowledge is most relevant to the incoming domain, and those mined prompts initialize the new prompt set to create an initial alignment; an online alignment loss then keeps the new prompts' components consistent with matched old prompts as training proceeds. The result is a DIL method that improves average accuracy over C-Prompt by 4.25, 4.08, 5.59, and 5.00 points on DomainNet, ImageNet-R, ImageNet-C, and ImageNet-Mix respectively, with no extra learnable parameters.

What carries the argument

KA-Prompt has two mechanisms. Reusable Knowledge Mining ($\Psi_M$) computes a relation matrix between historical prompt keys and new-domain image features, then greedily adds prompts to a reusable memory by repeatedly picking the prompt that contributes the most uncovered new-domain knowledge, measured by a cumulative score histogram. Aligning-guided New Prompt Learning ($\Psi_L$) initializes the new prompt set from that mined memory, then during training matches old prompts to the current input, weights each old prompt by how much its similarity falls below the new prompts' minimal matched score, and fuses old and new prompts into an aligned prompt whose cross-entropy loss $\mathcal{L}_{agn}$ is added to the ordinary classification loss. The greedy selection supplies the initial component correspondence; the online loss is meant to preserve that correspondence as new prompts drift during training.

What would settle it

Run KA-Prompt to completion, then at inference apply a single fixed random permutation to the component positions of every prompt before fusion; the paper's own ablation shows accuracy drops under shuffling, so if the drop disappeared, or a shuffled variant matched or beat the unshuffled model, the claim that aligned component positions drive the gains would be falsified.

Watch

Extended reading notes

Core claim

The paper's central claim is that prompts are not monolithic: each prompt's $L_p$ components encode distinct types of knowledge, such as object-part features, and for prompts learned on different domains the same kind of knowledge tends to land in different component slots. When inference fuses the top-$K$ matched prompts by averaging, misaligned components merge unrelated knowledge into one vector, producing interference; random shuffling of component positions before fusion changes C-Prompt's accuracy, which the paper reads as evidence that the learned component order is far from optimal. KA-Prompt therefore enforces component-wise alignment: reusable old prompts are selected greedily by how much new-domain knowledge they uniquely cover, new prompts are initialized from them, and a weighted old-new fusion with a cross-entropy loss acts as an alignment-preserving regularizer while the new prompts learn. The paper reports that this yields consistent gains over C-Prompt on four DIL benchmarks, with the largest gains appearing in later domains as shared knowledge accumulates.

Load-bearing premise

The load-bearing assumption is that training on a weighted average of old and new prompts forces matching knowledge components into matching positions, even though the method never directly constrains or measures which component in a new prompt corresponds to which component in an old prompt.

Editorial extensions

If this is right

  • If component alignment is the cause of the gains, then any prompt-based DIL method that fuses prompts across domains should benefit from a similar initialization-from-mined-prompts step, not just KA-Prompt.
  • Reusable knowledge can be transferred across domains without storing raw images, relying only on prompt keys and pre-trained features, which matters for privacy-sensitive applications.
  • Because no learnable parameters are added, the accuracy gains come from how prompts are initialized and regularized, not from a larger model.
  • The method's advantage grows with the number of seen domains, suggesting that alignment converts accumulated old knowledge into a reusable asset rather than a source of interference.

Reading between the lines

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

  • Editorial inference: the alignment loss is an indirect proxy; a direct component-correspondence metric, such as tracking how far each new component drifts from the old component it was initialized from, would isolate whether the mechanism is really alignment or generic regularization.
  • Editorial inference: the greedy coverage objective resembles submodular maximization, so formalizing prompt selection that way could yield approximation guarantees and transfer to other memory-and-retrieval selection problems.
  • Editorial inference: the same alignment principle might extend beyond DIL to task-incremental or class-incremental prompt pools, where reusable modules are also trained independently and fused later.
  • Editorial inference: a testable extension is to vary domain order so that some adjacent domains share little knowledge; if gains shrink, the mining step is doing the work, while persistent gains would point to the online alignment loss as a stronger regularizer.
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 / 5 minor

Summary. The paper introduces KA-Prompt, a prompt-based domain incremental learning method built on the observation that per-domain prompts learned independently suffer from component-wise misalignment, so that fusing prompts across domains mixes unrelated knowledge components. To address this, the method (1) mines a set of relevant historical prompts via a greedy search and uses them to initialize new prompts, and (2) applies an online 'alignment' loss that trains the new prompts on a weighted elementwise fusion of old and new prompts. The method is evaluated on DomainNet, ImageNet-R, ImageNet-C, and ImageNet-Mix, reporting consistent gains over C-Prompt and other baselines, with ablations of the proposed modules, hyperparameter sensitivity, and overhead comparisons. The code is publicly available.

Significance. The paper addresses a real and practically relevant problem in prompt-based domain incremental learning, and the reported empirical gains are substantial and consistent across four benchmarks. The strengths of the manuscript are its thorough experimental evaluation, including per-domain results, module ablations, hyperparameter studies, and cost comparisons, as well as the release of source code. The proposed reusable-knowledge mining idea is interesting and could be useful beyond the specific alignment claim. However, the paper's central causal claim — that the gains come from component-wise prompt-knowledge alignment — rests on an indirect proxy: the alignment loss in Eq. (10) is a cross-entropy loss on a weighted elementwise fusion, and the manuscript does not directly measure or constrain component correspondence. The shuffle evidence is suggestive but not conclusive, because any model trained with an elementwise fusion auxiliary loss will be sensitive to component permutation. If the mechanism attribution is correctly established by additional experiments, the contribution would be solid; as written, the central mechanism remains undersupported.

major comments (3)
  1. [Section 3.3, Eq. (10)] The alignment loss L_agn is a cross-entropy loss on p_agn^x, which is an elementwise weighted average of old and new prompts. This loss does not identify, constrain, or measure which component slot carries which knowledge; minimizing it could simply make the new prompt statistically compatible with the old prompts in aggregate. Thus Eq. (10) functions as a fusion regularizer or ensemble objective, and the paper's core claim that it establishes componential knowledge alignment is not directly supported. Please provide a direct measurement of component-wise correspondence (e.g., a matching or correspondence metric between prompt components before and after training, or an analysis that component permutations destroy the learned structure in a way that a generic regularizer would not), or add a control experiment using an equally strong fusion regularizer that does not rely on specific component positions.
  2. [Section 4.3, Fig. 7] The prompt-shuffle ablation is indirect and confounded. A model trained with an auxiliary loss on a specific elementwise fusion scheme will necessarily be sensitive to shuffling components under that scheme, even if no semantic alignment was ever established. The comparison with C-Prompt does not isolate the alignment mechanism, because C-Prompt is not trained with any elementwise fusion loss. A stronger test would be to compare KA-Prompt with a baseline that uses the same online fusion loss and hyperparameters but without the mined initialization, or to shuffle according to the mined correspondence versus random permutations and show that the alignment-consistent ordering is uniquely the best.
  3. [Appendix A] The fallback branch in the search algorithm, which randomly interpolates R=2 prompts when the cumulative score histogram is zero, is uncontrolled. The manuscript does not report how often this branch is taken, for which domains, or how the random interpolation affects the reported averages. Since the mined initialization is one of the two main mechanisms, this branch needs a frequency analysis and an ablation with a deterministic fallback (e.g., nearest prompt or a learned interpolation) to show that the random choice is not influencing the results.
minor comments (5)
  1. [Section 3.3, heading] The subsection heading 'Reusable Knowledge Ming' is a typo; it should be 'Reusable Knowledge Mining'.
  2. [Section 4.1] The word 'hypermeters' should be 'hyperparameters'.
  3. [Fig. 7] Both subfigures in Fig. 7 are labeled '(a)'; the second should be labeled '(b)'.
  4. [Algorithm 2] In the line for obtaining top-K old prompts, the notation uses indices for the new prompt set (p_i^t) instead of the historical prompt set (p_i^{t-1}); please correct the indexing.
  5. [Equation (4) and surrounding text] The phrase 'An base relation matrix' should be 'A base relation matrix', and the dimensions in the displayed equation should be checked for consistency with the definition of S0.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: KA-Prompt's contribution is an empirically evaluated heuristic; no derivation step reduces to its own inputs or to a load-bearing self-citation.

full rationale

The paper's central claims are empirical: component-wise prompt misalignment causes interference during fusion, and the proposed mining plus alignment losses improve accuracy on four benchmarks. I checked the derivation chain for reductions of the kind the circularity pass targets. The alignment loss in Eq. 10 is a cross-entropy loss applied to the weighted fusion of old and new prompts in Eq. 9. This is a heuristic proxy for alignment, not a quantity that is defined in terms of the reported accuracy, and it is not a fitted parameter renamed as a prediction. The reported Avg-ACC numbers are measured on held-out test domains and compared against external baselines, including C-Prompt; they are not outputs of an equation whose constants were fitted to those test sets. The greedily mined prompt memory and the online alignment loss are trained with classification objectives; neither is constructed to encode the final benchmark numbers. Hyperparameters tau and lambda are tuned, but tuning auxiliary weights is standard practice and does not make the benchmark comparison a fitted prediction. Self-citations are present, chiefly to C-Prompt (Liu et al., 2024a), which is also from the corresponding author's group. However, C-Prompt is used as a baseline and as the starting framework, not as an unverified authority that forces the paper's conclusions; the paper's novelty is a training-time modification evaluated against that baseline on independent benchmarks. There is no uniqueness theorem imported from the authors' prior work, no ansatz smuggled in via citation, and no known result merely renamed. The apparent weakness that the alignment loss does not directly constrain component correspondence is a correctness or validity concern about the heuristic, not circularity: the loss is not equivalent by construction to the outcome it is claimed to produce. Therefore the honest finding is no significant circularity, score 0.

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

The method depends on three main assumptions beyond standard ML practice: prompt components have positionally meaningful semantics (Sec. 3.2), cosine similarity between keys and features measures reusable knowledge (Eq. 4), and the fusion-based alignment loss preserves the desired alignment (Eq. 10). tau and lambda are tuned on ImageNet-R (Fig. 6) and used as defaults elsewhere; the Appendix-A interpolation fallback adds another hand-chosen element. No new entities are introduced.

free parameters (3)
  • tau = 0.01
    Temperature scaling old-prompt weights in Eq. 8; tuned on ImageNet-R (Fig. 6).
  • lambda = 0.1
    Weight of alignment loss in Eq. 11; tuned on ImageNet-R (Fig. 6).
  • R = 2
    Number of prompts interpolated when the cumulative score histogram is all zero (Appendix A); ad hoc, may affect a small number of tasks.
assumptions (4)
  • domain assumption Prompt components at the same positional index encode comparable knowledge across domains.
    Sec. 3.2 and Fig. 2a; the entire alignment design relies on this.
  • domain assumption Cosine similarity between prompt keys and new-domain image features (Eq. 4) measures reusable knowledge.
    Sec. 3.3, fR; if this metric is poor, greedy mining will select irrelevant prompts.
  • ad hoc to paper The weighted fusion of old and new prompts in Eq. 9, trained with CE on new labels, preserves alignment.
    Sec. 3.3, online aligning; implicit regularizer, no explicit component-matching term.
  • ad hoc to paper When the cumulative score histogram is zero, interpolating two randomly chosen old prompts (Appendix A) is a harmless fallback.
    Appendix A note; only used in edge cases but could inject noise.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Componential Prompt-Knowledge Alignment for Domain Incremental Learning." pith.science (2026). https://pith.science/paper/RMIKU23K

@misc{pith2026250504575,
  author       = {Pith},
  title        = {Pith review of: Componential Prompt-Knowledge Alignment for Domain Incremental Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RMIKU23K}},
  note         = {Machine review of arXiv:2505.04575}
}
read the original abstract

Domain Incremental Learning (DIL) aims to learn from non-stationary data streams across domains while retaining and utilizing past knowledge. Although prompt-based methods effectively store multi-domain knowledge in prompt parameters and obtain advanced performance through cross-domain prompt fusion, we reveal an intrinsic limitation: component-wise misalignment between domain-specific prompts leads to conflicting knowledge integration and degraded predictions. This arises from the random positioning of knowledge components within prompts, where irrelevant component fusion introduces interference.To address this, we propose Componential Prompt-Knowledge Alignment (KA-Prompt), a novel prompt-based DIL method that introduces component-aware prompt-knowledge alignment during training, significantly improving both the learning and inference capacity of the model. KA-Prompt operates in two phases: (1) Initial Componential Structure Configuring, where a set of old prompts containing knowledge relevant to the new domain are mined via greedy search, which is then exploited to initialize new prompts to achieve reusable knowledge transfer and establish intrinsic alignment between new and old prompts. (2) Online Alignment Preservation, which dynamically identifies the target old prompts and applies adaptive componential consistency constraints as new prompts evolve. Extensive experiments on DIL benchmarks demonstrate the effectiveness of our KA-Prompt. Our source code is available at https://github.com/zhoujiahuan1991/ICML2025-KA-Prompt

Figures

Figures reproduced from arXiv: 2505.04575 by the authors.

Figure 1
Figure 1. (a) DIL aims to learn with a stream of data [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. (a) Within a prompt, different components typi [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The illustration of our KA-Prompt method. When the new domain data [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: The seen domain performance tendency along domain incremental learning process. [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 6
Figure 6. Figure 6: Ablation studies on the hyper-parameters under [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: Performance comparison on the last domain of [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 8
Figure 8. Figure 8: The seen domain performance tendency on DomainNet. [PITH_FULL_IMAGE:figures/full_fig_p013_8.png]
Figure 9
Figure 9. Figure 9: The seen domain performance tendency on ImageNet-Mix. [PITH_FULL_IMAGE:figures/full_fig_p013_9.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. UPP: Unified Point-Level Prompting for Robust Point Cloud Analysis

    cs.CV 2025-07 conditional novelty 6.0 of 10

    UPP unifies denoising and completion as point-level prompts in a frozen pre-trained point cloud transformer, improving classification accuracy on noisy and incomplete point clouds without updating backbone weights.

Reference graph

Works this paper leans on

34 extracted references · 33 canonical work pages · cited by 1 Pith paper

  1. [1]

    Prototype-sample relation distillation: towards replay-free continual learning

    Belilovsky, E. Prototype-sample relation distillation: towards replay-free continual learning. In Interna- tional Conference on Machine Learning, pp. 1093–1106. PMLR, 2023

  2. [2]

    Mind: Multi-task incremental network distillation

    Bonato, J., Pelosin, F., Sabetta, L., and Nicolosi, A. Mind: Multi-task incremental network distillation. In Proceed- ings of the AAAI Conference on Artificial Intelligence , volume 38, pp. 11105–11113, 2024

  3. [3]

    On the stability-plasticity dilemma in continual meta-learning: theory and algorithm

    Chen, Q., Shui, C., Han, L., and Marchand, M. On the stability-plasticity dilemma in continual meta-learning: theory and algorithm. Advances in Neural Information Processing Systems, 36, 2024

  4. [4]

    Cp-prompt: Composition-based cross-modal prompting for domain-incremental continual learning

    Song, M. Cp-prompt: Composition-based cross-modal prompting for domain-incremental continual learning. In Proceedings of the 32nd ACM International Conference on Multimedia, pp. 2729–2738, 2024

  5. [5]

    Fini, E., Da Costa, V . G. T., Alameda-Pineda, X., Ricci, E., Alahari, K., and Mairal, J. Self-supervised models are continual learners. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 9621–9630, 2022

  6. [6]

    Greedy poisson rejection sampling

    Flamich, G. Greedy poisson rejection sampling. Advances in Neural Information Processing Systems, 36, 2024

  7. [7]

    Consistent prompting for rehearsal-free continual learning

    Gao, Z., Cen, J., and Chang, X. Consistent prompting for rehearsal-free continual learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 28463–28473, 2024

  8. [8]

    Person re-identification method based on color attack and joint defence

    Gong, Y ., Huang, L., and Chen, L. Person re-identification method based on color attack and joint defence. In Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp. 4313–4322, 2022

Show all 34 references
  1. [9]

    and Dietterich, T

    Hendrycks, D. and Dietterich, T. G. Benchmarking neural network robustness to common corruptions and surface variations. arXiv preprint arXiv:1807.01697, 2018. 9 Componential Prompt-Knowledge Alignment for Domain Incremental Learning

  2. [10]

    The many faces of robustness: A critical analysis of out-of-distribution generalization

    Dorundo, E., Desai, R., Zhu, T., Parajuli, S., Guo, M., et al. The many faces of robustness: A critical analysis of out-of-distribution generalization. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 8340–8349, 2021

  3. [11]

    Gradual di- vergence for seamless adaptation: A novel domain in- cremental learning method

    Jeeveswaran, K., Arani, E., and Zonooz, B. Gradual di- vergence for seamless adaptation: A novel domain in- cremental learning method. In Proceedings of the 41st International Conference on Machine Learning, volume 235, pp. 21486–21501. PMLR, 2024

  4. [12]

    and Choi, D.-W

    Kang, H. and Choi, D.-W. Recall-oriented continual learn- ing with generative adversarial meta-model. In Proceed- ings of the AAAI Conference on Artificial Intelligence , volume 38, pp. 13040–13048, 2024

  5. [13]

    Overcoming catastrophic forgetting in neural networks

    Grabska-Barwinska, A., et al. Overcoming catastrophic forgetting in neural networks. Proceedings of the national academy of sciences, 114(13):3521–3526, 2017

  6. [14]

    Clustering-based domain- incremental learning

    Lamers, C., Vidal, R., Belbachir, N., van Stein, N., B¨aeck, T., and Giampouras, P. Clustering-based domain- incremental learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 3384– 3392, 2023

  7. [15]

    Person- alized federated domain-incremental learning based on adaptive knowledge matching

    Li, Y ., Xu, W., Wang, H., Qi, Y ., Guo, J., and Li, R. Person- alized federated domain-incremental learning based on adaptive knowledge matching. In European Conference on Computer Vision, pp. 127–144. Springer, 2025

  8. [16]

    and Hoiem, D

    Li, Z. and Hoiem, D. Learning without forgetting. IEEE Transactions on Pattern Analysis and Machine Intelli- gence, 40(12):2935–2947, 2017

  9. [17]

    and Li, W.-J

    Liang, Y .-S. and Li, W.-J. Inflora: Interference-free low- rank adaptation for continual learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 23638–23647, 2024

  10. [18]

    Stop: Integrated spatial-temporal dynamic prompting for video understanding

    Liu, Z., Xu, K., Su, B., Zou, X., Peng, Y ., and Zhou, J. Stop: Integrated spatial-temporal dynamic prompting for video understanding. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2025

  11. [19]

    J., Masana, M., Possegger, H., and Bischof, H

    Mirza, M. J., Masana, M., Possegger, H., and Bischof, H. An efficient domain-incremental learning approach to drive in all weather conditions. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 3001–3011, 2022

  12. [20]

    and Wolf, L

    Oren, G. and Wolf, L. In defense of the learning without forgetting for task incremental learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 2209–2218, 2021

  13. [21]

    Moment matching for multi-source domain adaptation

    Wang, B. Moment matching for multi-source domain adaptation. In Proceedings of the IEEE/CVF Interna- tional Conference on Computer Vision, pp. 1406–1415, 2019

  14. [22]

    Learning prompt- enhanced context features for weakly-supervised video anomaly detection

    Pu, Y ., Wu, X., Yang, L., and Wang, S. Learning prompt- enhanced context features for weakly-supervised video anomaly detection. IEEE Transactions on Image Process- ing, 2024

  15. [23]

    Learning transferable visual models from natural language supervision

    Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., et al. Learning transferable visual models from natural language supervision. In International Conference on Machine Learning, pp. 8748–8763. PMLR, 2021

  16. [24]

    Vision transformers with mixed-resolution tokenization

    Ronen, T., Levy, O., and Golbert, A. Vision transformers with mixed-resolution tokenization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 4613–4622, 2023

  17. [25]

    and Wang, H

    Shi, H. and Wang, H. A unified approach to domain in- cremental learning with memory: Theory and algorithm. Advances in Neural Information Processing Systems, 36, 2024. 10 Componential Prompt-Knowledge Alignment for Domain Incremental Learning

  18. [26]

    Coda-prompt: Continual decomposed attention-based prompting for rehearsal-free continual learning

    Kim, D., Arbelle, A., Panda, R., Feris, R., and Kira, Z. Coda-prompt: Continual decomposed attention-based prompting for rehearsal-free continual learning. In Pro- ceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pp. 11909–11919, 2023

  19. [27]

    Non-exemplar domain incremental object detection via learning domain bias

    Song, X., He, Y ., Dong, S., and Gong, Y . Non-exemplar domain incremental object detection via learning domain bias. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pp. 15056–15065, 2024. Van de Ven, G. M., Tuytelaars, T., and Tolias, A. S. Three ...

  20. [28]

    Non-exemplar domain incremental learning via cross- domain concept integration

    Wang, Q., He, Y ., Dong, S., Gao, X., Wang, S., and Gong, Y . Non-exemplar domain incremental learning via cross- domain concept integration. In European Conference on Computer Vision, pp. 144–162. Springer, 2025

  21. [29]

    Not all images are worth 16x16 words: Dynamic transformers for efficient image recognition

    Wang, Y ., Huang, R., Song, S., Huang, Z., and Huang, G. Not all images are worth 16x16 words: Dynamic transformers for efficient image recognition. Advances in Neural Information Processing Systems , 34:11960– 11973, 2021

  22. [30]

    General incremental learn- ing with domain-aware categorical representations

    Xie, J., Yan, S., and He, X. General incremental learn- ing with domain-aware categorical representations. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 14351–14360, 2022

  23. [31]

    Dask: Distribution rehearsing via adaptive style kernel learning for exemplar-free lifelong person re-identification

    Xu, K., Jiang, C., Xiong, P., Peng, Y ., and Zhou, J. Dask: Distribution rehearsing via adaptive style kernel learning for exemplar-free lifelong person re-identification. In Proceedings of the AAAI Conference on Artificial Intelli- gence, volume 39, pp. 8915–8923, 2025

  24. [32]

    Mma-diffusion: Multimodal attack on diffusion models

    Yang, Y ., Gao, R., Wang, X., Ho, T.-Y ., Xu, N., and Xu, Q. Mma-diffusion: Multimodal attack on diffusion models. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition , pp. 7737–7746, 2024

  25. [33]

    Scap: Transductive test-time adaptation via supportive clique-based attribute prompting

    Zhang, C., Xu, K., Liu, Z., Peng, Y ., and Zhou, J. Scap: Transductive test-time adaptation via supportive clique-based attribute prompting. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2025

  26. [34]

    Revisiting class-incremental learning with pre- trained models: Generalizability and adaptivity are all you need

    Liu, Z. Revisiting class-incremental learning with pre- trained models: Generalizability and adaptivity are all you need. International Journal of Computer Vision, pp. 1–21, 2024. 11 Componential Prompt-Knowledge Alignment for Domain Incremental Learning A. Algorithm. The over...

Pith tools

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