Pith. sign in

REVIEW 4 major objections 4 minor 48 references

AnchorInv: Few-Shot Class-Incremental Learning of Physiological Signals via Representation Space Guided Inversion

T0 review · 4 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read AnchorInv stores feature-space anchor points instead of raw signals, inverts them into replay samples, and reports state-of-the-art few-shot incremental learning results on three EEG and EMG datasets.

desk verdict Solid empirical FSCIL paper with a useful anchor-inversion trick and a good multi-trial protocol, but the 'faithful replay' claim needs more evidence and the finetune tuning is ambiguous. read the letter →

arxiv 2412.13714 v1 pith:SFJDQCXG submitted 2024-12-18 cs.LG

Chenqi Li , Boyan Gao , Gabriel Jones , Timothy Denison , Tingting Zhu This is my paper · ORCID
classification cs.LG
keywords few-shotclass-incrementallearningmodelinversionphysiologicaltimeseriesEEGEMGcatastrophicforgettingbufferreplayfeaturespace
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

This paper sets out to establish that a replay-based few-shot class-incremental learning method can work for physiological time series even when the base training set has very few classes. Instead of storing raw signals, AnchorInv stores a small set of anchor points in the feature space, synthesizes replay samples by inverting those anchors, and finetunes the backbone on the synthetic samples together with the new few-shot examples. The paper argues that this preserves old-class knowledge while adapting to new classes and respects privacy, and it reports that the method outperforms state-of-the-art baselines across three EEG and EMG datasets. The paper also argues that existing FSCIL evaluations are too sensitive to which few-shot samples happen to be drawn, so it repeats adaptation over many randomly sampled support sets to report mean and standard deviation.

What carries the argument

The carrying mechanism is the anchor set: a small collection of embedding vectors, one or more per base class, selected by random sampling or clustering from the feature distribution of the training set. At each incremental session these anchors are the inversion targets; minimizing mean absolute error between a randomly initialized input's embedding and an anchor yields a synthetic replay sample, and the replay set is used in a finetuning loss that regularizes old-class knowledge while the backbone adapts to new classes. The metric classifier is a prototype-based softmax over negative cosine distances, so keeping old-class embeddings near their anchors is what prevents the feature space from shifting when new classes are integrated.

What would settle it

Measure the last-session Macro-F1 on NHIE and GRABMyo when the finetuning regularization uses real stored samples instead of inverted ones, a real-replay upper bound; if AnchorInv falls clearly below that bound on either dataset, the claim that inverted samples carry as much prior-class information as real samples is contradicted. A second check: replace the anchor-guided mean-absolute-error inversion loss with a label-space cross-entropy inversion loss; if performance does not drop, feature-space guidance specifically is not the cause of the reported gains.

Watch

Extended reading notes

Core claim

The central claim is that guiding model inversion by feature-space anchor points, rather than by label-space classification loss, produces replay samples that prevent catastrophic forgetting more effectively during few-shot incremental sessions. Concretely, after base training the method computes a set of anchor points summarizing each base class's embeddings, solves an optimization to synthesize samples whose embeddings match those anchors under mean absolute error, and then finetunes using a loss that combines cross-entropy on the new few-shot samples with a regularization term on the inverted old-class samples. The paper reports that this achieves the best Macro-F1 across all incremental sessions on BCI, NHIE, and GRABMyo, with statistically significant differences from baselines at the p < 0.05 level by a two-sided signed-rank test, and that its performance closely tracks a real-replay upper bound on BCI.

Load-bearing premise

The load-bearing premise is that making a synthetic signal map to the same internal representation as an old class also makes that synthetic signal useful for preserving the old class during later training, even if the synthetic signal does not look like a real one.

Editorial extensions

If this is right

  • Incremental learning on physiological signals can be privacy-preserving: no raw base-session samples need to be stored, only anchor points in feature space.
  • The method is most valuable exactly when base classes are scarce, because it finetunes the backbone instead of relying on a frozen generalizable feature extractor.
  • Reporting mean and standard deviation across multiple randomly sampled few-shot support sets gives a more stable expectation of real-world performance than a single split.
  • On BCI, AnchorInv's final performance is close to that of replaying real stored samples, suggesting the synthetic samples carry comparable regularization information.

Reading between the lines

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

  • Because inversion only needs the embedding network and stored anchors, the same procedure should transfer to other physiological modalities or restricted-data settings; testing on ECG or wearable motion signals would be a natural next step.
  • The visual distinctness of inverted samples from real signals suggests AnchorInv may be regularizing the feature space rather than reproducing faithful waveforms, meaning its success does not depend on replay samples looking realistic.
  • The paper only tests the real-replay upper bound on BCI; replicating that comparison on NHIE and GRABMyo would show whether the near-RealReplay result generalizes or is specific to within-subject BCI data.
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 / 4 minor

Summary. The paper proposes AnchorInv, a buffer-replay method for few-shot class-incremental learning (FSCIL) on physiological time series. Instead of storing raw samples, AnchorInv stores selected feature-space anchor points from each session and uses model inversion—minimizing mean absolute error between the embedding of a synthetic input and a stored anchor (Eq. 11)—to generate replay samples. These synthetic samples are then used together with the new few-shot samples to finetune the backbone (Eqs. 12–14). The authors evaluate on three public datasets (BCI-IV 2a, NHIE, GRABMyo) with 1-way-10-shot incremental sessions, reporting Macro-F1 across multiple random trials. They report that AnchorInv achieves the best Macro-F1 on all datasets across all sessions, with statistically significant differences from baselines, and present ablations on anchor selection, number of anchors, number of shots, and base-class count. The paper also claims the approach protects data privacy because raw data are not stored.

Significance. The paper targets a real and under-studied setting: FSCIL for physiological signals with very few base classes, where the authors argue that backbone finetuning is necessary. The proposed idea—using feature-space anchors to guide inversion for replay—is reasonable and distinct from label-space inversion baselines such as DeepDream and DeepInv. The manuscript has several strengths: it performs multi-trial evaluation with reported means and standard deviations, applies Wilcoxon signed-rank tests, compares against a range of baselines, and includes ablations on anchor selection, anchor count, shot count, and base-class count. If the central mechanism is verified, the approach would be a practical contribution to privacy-conscious incremental learning for biomedical data. However, the paper's core evidence that the inverted samples are representative of old classes is incomplete, and there are unresolved questions about the fairness of the hyperparameter selection procedure relative to baselines.

major comments (4)
  1. [§4.3, Appendix E.2 (Table 8)] The RealReplay upper bound—the direct check that synthetic replay samples carry information comparable to real samples—is reported only for BCI (Table 8). For NHIE and GRABMyo, where the paper claims state-of-the-art results (Tables 2b and 2c), no such verification is provided. Since the central claim is that AnchorInv preserves prior knowledge by generating representative replay samples, the absence of this check leaves open the alternative explanation that the improvements arise from a generic regularizer rather than from faithful replay. Please provide RealReplay comparisons for NHIE and GRABMyo, or otherwise experimentally demonstrate that finetuning on the inverted samples transfers to held-out real data of the base classes.
  2. [§C.2 and Appendix D] The baseline finetuning schedules are explicitly selected on the test set: Appendix D states that the number of iterations is chosen by running 10 trials and selecting the best-performing iteration based on test-set performance. In contrast, Section C.2 reports AnchorInv's finetuning iteration ranges (e.g., 1000–1550 for BCI) without stating whether these were selected on the test set or by a validation criterion. If AnchorInv's schedule was tuned on test performance while baseline schedules were chosen in the same way, this should be stated; if AnchorInv's schedule was chosen differently, the comparison may be unfair. Please clarify the selection procedure and, if needed, report results under a unified validation protocol.
  3. [§3.2, Eq. (11), Figures 3 and 5] The qualitative evidence in Figure 3 that inverted feature vectors overlap with anchors is guaranteed by the MAE objective in Eq. (11) and therefore cannot independently support the representativeness of the inverted samples. The paper itself notes in Section 4.3 (Qualitative Analysis) and Figure 5 that real and inverted samples are visually distinct and that inverted samples are dominated by high-frequency fluctuations. This suggests the inversion may find off-manifold inputs that match the anchor embedding without lying on the true data manifold. Please add a quantitative test of representativeness, for example: (a) train a classifier on inverted samples and evaluate on real base-class test data, or (b) measure whether finetuning with Lold on inverted samples changes the model's embeddings of real base-class inputs in the intended way. Such a test would directly support the claim that the replay samples are 'faithful and representative' rather than merely matching the anchors in feature space.
  4. [Abstract, §1, §3.2] The claim that AnchorInv 'protects privacy' and 'adheres to data privacy and sharing regulations' is asserted without a formal privacy analysis or a well-defined threat model. Storing feature-space anchors and generating synthetic samples from them reduces the direct exposure of raw data, but the paper demonstrates that anchors can be inverted to produce samples, so the privacy guarantee is not automatic. Please either provide a formal statement (e.g., under what adversary model the stored anchors cannot be used to recover identifiable raw signals) or soften the claim to a practical mitigation rather than a guarantee. This is load-bearing because privacy preservation is stated as a core motivation and contribution.
minor comments (4)
  1. [Appendix H and Reproducibility Checklist] The reproducibility checklist states that all source code implementing the new method is included and will be made publicly available, while Appendix H says only that code will be made available upon publication. Please align these statements so the reader knows what is available at submission time.
  2. [Table 7 and Appendix G] There are typographical errors: 'GRAMyo' in the Table 7 caption should be 'GRABMyo', and 'Distrbution' in the captions of Figures 7 and 8 should be 'Distribution'.
  3. [§4.3 and Table 2] The sentence 'Across all incremental sessions, AnchorInv is statistically significantly different from all baseline methods' is ambiguous because it does not state the pairing used in the Wilcoxon test (e.g., paired per trial and session) or whether a multiple-comparison correction was applied. Please clarify the test protocol.
  4. [§C.1 and §C.2] The base-session performance differs among methods because Finetune and NC-FSCIL use an MLP classifier while the other methods use a metric classifier; the paper explains this, but the base-session rows in Table 2 are not directly comparable. It would be helpful to state this explicitly in the table caption or the main text.

Circularity Check

1 steps flagged · score 2.0 of 10

AnchorInv's main results are an external empirical benchmark; the only circular element is using the inversion objective's own convergence as evidence of 'faithful representation.'

  1. self definitional [Section 4.3, Qualitative Analysis of Inversion (Figure 3); Eq. 11]
    "Note that the inverted feature vector and the anchor points overlap. This shows that the proposed model inversion algorithm can synthesize inverted samples that project to any target anchor point in the feature space."

    The overlap is guaranteed by the inversion definition: Eq. 11 sets the replay sample as the argmin over x-hat of MAE between f_theta(x-hat) and the anchor a, and Appendix C.2 optimizes this objective for thousands of iterations. Therefore any successful inversion will by construction have an embedding close to the anchor; Figure 3 is a sanity check that the optimizer converged, not independent evidence that the generated samples are 'representative' or 'faithful.' The paper's own Figure 5 confirms real and inverted EEG are visually distinct despite feature-level MAE of only 0.06-0.08. Using this overlap as qualitative support for the 'faithful representation' mechanism is thus circular: the evidence is built into the objective being optimized.

full rationale

The central claims of AnchorInv are empirical benchmark results against external baselines (Finetune, NC-FSCIL, ProtoNet, TEEN, DeepDream, DeepInv) on public datasets (BCI-IV 2a, NHIE, GRABMyo), with repeated trials and Wilcoxon signed-rank tests. Those results do not reduce to a fitted parameter renamed as a prediction, and the paper does not rely on a self-citation chain or an imported uniqueness theorem. The proposed pipeline (Eqs. 1-14) is a well-defined construction, and the reported Macro-F1 scores are independent of the method's own assumptions in the sense that they are measured against held-out test data. The RealReplay upper bound on BCI (Table 8) provides external grounding for the faithfulness mechanism, though it is not reported for NHIE and GRABMyo; that absence is an evidence gap about generality, not circularity. The one circular step is the qualitative claim in Figure 3, where the overlap between inverted features and anchors is presented as evidence for the method's representational faithfulness, even though that overlap is literally the loss minimized in Eq. 11. Because this tautological evidence is not the basis of the main SOTA comparison, the overall circularity is minor rather than structural.

Assumptions & free parameters 5 free parameters · 3 assumptions · 0 invented entities

The method introduces algorithmic constructs, anchor points and inverted replay samples, but no new physical entities or forces. The key assumptions are about the informativeness of feature-space anchors, the representativeness of inverted samples, and privacy. Many hyperparameters, such as lambda, P, inversion iterations, and finetuning lengths, are chosen by hand or per dataset.

free parameters (5)
  • lambda (replay loss weight) = not reported
    Controls the balance between new-class loss and old-class replay loss in Eq. 12 and is never given a value in the paper.
  • P (anchor points per class) = 50
    Number of feature anchors stored per base class, chosen by hand; ablation studies vary this value.
  • Inversion iterations = 4000 (BCI, NHIE), 2000 (GRABMyo)
    Number of gradient steps for synthesizing each replay sample, chosen per dataset.
  • Finetuning iterations = 1000-1550 (BCI), 700-1250 (NHIE), 0-1300 (GRABMyo)
    Dataset-dependent finetune lengths; the paper does not state whether these were selected on the test set, while baseline schedules explicitly were.
  • Classifier temperature T = 16
    Temperature in the cosine classifier (Eq. 4).
assumptions (3)
  • domain assumption Minimizing MAE between a synthetic input's embedding and a stored anchor point yields replay samples representative of the true class distribution.
    Core premise of the inversion step (Eq. 11); if false, replay samples are off-manifold and Lold may distort old knowledge. Qualitative analysis in Sec. 4.3 provides only weak support.
  • domain assumption Anchor points randomly sampled from base-class features summarize the class-conditional distribution well enough for replay.
    Used in Appendix C.1 with M=50 random anchors; ablations show dataset-dependent sensitivity.
  • ad hoc to paper Storing feature-space anchors instead of raw data protects data privacy.
    Asserted in the Abstract and Sec. 1; feature vectors can still encode sensitive information and model inversion is a known privacy attack, so this assumption is unsupported.

how reviews work

0 comments
Cite this review

Pith. "Pith review of AnchorInv: Few-Shot Class-Incremental Learning of Physiological Signals via Representation Space Guided Inversion." pith.science (2026). https://pith.science/paper/SFJDQCXG

@misc{pith2026241213714,
  author       = {Pith},
  title        = {Pith review of: AnchorInv: Few-Shot Class-Incremental Learning of Physiological Signals via Representation Space Guided Inversion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SFJDQCXG}},
  note         = {Machine review of arXiv:2412.13714}
}
read the original abstract

Deep learning models have demonstrated exceptional performance in a variety of real-world applications. These successes are often attributed to strong base models that can generalize to novel tasks with limited supporting data while keeping prior knowledge intact. However, these impressive results are based on the availability of a large amount of high-quality data, which is often lacking in specialized biomedical applications. In such fields, models are usually developed with limited data that arrive incrementally with novel categories. This requires the model to adapt to new information while preserving existing knowledge. Few-Shot Class-Incremental Learning (FSCIL) methods offer a promising approach to addressing these challenges, but they also depend on strong base models that face the same aforementioned limitations. To overcome these constraints, we propose AnchorInv following the straightforward and efficient buffer-replay strategy. Instead of selecting and storing raw data, AnchorInv generates synthetic samples guided by anchor points in the feature space. This approach protects privacy and regularizes the model for adaptation. When evaluated on three public physiological time series datasets, AnchorInv exhibits efficient knowledge forgetting prevention and improved adaptation to novel classes, surpassing state-of-the-art baselines.

Figures

Figures reproduced from arXiv: 2412.13714 by the authors.

Figure 1
Figure 1. Overview of AnchorInv. In the base session, the training dataset is projected to the feature space, and anchor points for each class are identified and saved in the anchor set memory. In incremental sessions, the anchor set guides the model inversion process to generate representative samples of previously seen classes. The inverted samples and the few-shot training set are subsequently used to finetune the backbone… view at source ↗
Figure 2
Figure 2. Overview of FSCIL Evaluation Procedure a) il￾lustrates the FSCIL evaluation process in the existing liter￾ature. b) illustrates the FSCIL evaluation process adopted in this work. In Session 1, we create M copies of the base session model. In each incremental session, each copy is adapted with a randomly sampled training set. Each copy is evaluated on the test set and the mean and standard devia￾tion are reported acr… view at source ↗
Figure 3
Figure 3. t-SNE Visualization of the Anchor Set and the Inverted Samples in Feature Space 4.4 Ablation Studies Selection of Anchor Points from Base Session We re￾peat AnchorInv experiments from Table 2a, 2b and 2c and study how different anchor selection algorithms for the base classes impact the last session performance. Instead of ran￾dom sampling, we select anchor points as feature vectors that are closest to class prototy… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: b shows that the performance stabilizes beyond 10 shots. Anchor Selection BCI NHIE GRABMyo Closest 38.65 ± 1.26 † 62.63±2.38 † 60.67±0.70 5 k-means Cluster 36.97±1.28 66.05±2.90 60.70±0.68 Random Closest 30% 38.36±1.34 † 63.07±2.39 † 60.67±0.71 † Random Closest 50% 38.…
Figure 5
Figure 5. Figure 5: Visualization of EEG Samples from NHIE (a) Anchor points selected as “Random Sample” (b) Anchor points selected through “Closest” to prototype [PITH_FULL_IMAGE:figures/full_fig_p014_5.png]
Figure 7
Figure 7. Figure 7: Visualization of Base Class Feature Space Dis￾trbution for the NHIE Dataset (a) Anchor points selected as “Random Sample” (b) Anchor points selected through “Closest” to prototype [PITH_FULL_IMAGE:figures/full_fig_p015_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 38 canonical work pages

  1. [1]

    Agarwal, A.; Banerjee, B.; Cuzzolin, F.; and Chaudhuri, S. 2022. Semantics-driven generative replay for few-shot class incremental learning. In Proceedings of the 30th ACM international conference on multimedia, 5246--5254

    work page 2022

  2. [2]

    u rek, A. F.; Aky \

    Aky \"u rek, A. F.; Aky \"u rek, E.; Wijaya, D. T.; and Andreas, J. 2021. Subspace regularizers for few-shot class incremental learning. arXiv preprint arXiv:2110.07059

    arXiv 2021

  3. [3]

    Arjovsky, M.; and Bottou, L. 2017. Towards principled methods for training generative adversarial networks. arXiv preprint arXiv:1701.04862

    arXiv 2017

  4. [4]

    u ller-Putz, G.; Schl \

    Brunner, C.; Leeb, R.; M \"u ller-Putz, G.; Schl \"o gl, A.; and Pfurtscheller, G. 2008. BCI Competition 2008--Graz data set A. Institute for knowledge discovery (laboratory of brain-computer interfaces), Graz University of Technology, 16: 1--6

    work page 2008

  5. [5]

    Chen, K.; and Lee, C.-G. 2021. Incremental few-shot learning via vector quantization in deep embedded space. In International conference on learning representations

    work page 2021

  6. [6]

    Chi, Z.; Gu, L.; Liu, H.; Wang, Y.; Yu, Y.; and Tang, J. 2022. Metafscil: A meta-learning approach for few-shot class incremental learning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 14166--14175

    work page 2022

  7. [7]

    Dong, S.; Hong, X.; Tao, X.; Chang, X.; Wei, X.; and Gong, Y. 2021. Few-shot class-incremental learning via relation knowledge distillation. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, 1255--1263

    work page 2021

  8. [8]

    Gao, Y.; Sharma, T.; and Cui, Y. 2023. Addressing the challenge of biomedical data inequality: An artificial intelligence perspective. Annual review of biomedical data science, 6(1): 153--171

    work page 2023

Show all 48 references
  1. [9]

    J.; Mirza, M.; Xiao, D.; Courville, A.; and Bengio, Y

    Goodfellow, I. J.; Mirza, M.; Xiao, D.; Courville, A.; and Bengio, Y. 2013. An empirical investigation of catastrophic forgetting in gradient-based neural networks. arXiv preprint arXiv:1312.6211

    2013 arXiv

  2. [10]

    Hersche, M.; Karunaratne, G.; Cherubini, G.; Benini, L.; Sebastian, A.; and Rahimi, A. 2022. Constrained few-shot class-incremental learning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 9057--9067

    2022

  3. [11]

    Jiang, N.; Pradhan, A.; and He, J. 2022. Gesture recognition and biometrics electroMyogram (GRABMyo). PhysioNet. doi: https://doi. org/10.13026/dzhr-nj10

    2022 doi

  4. [12]

    Kang, H.; Yoon, J.; Madjid, S. R. H.; Hwang, S. J.; and Yoo, C. D. 2023. On the Soft-Subnetwork for Few-shot Class Incremental Learning

    2023

  5. [13]

    Kim, D.-Y.; Han, D.-J.; Seo, J.; and Moon, J. 2023. WARPING THE SPACE : WEIGHT SPACE ROTATION FOR CLASS-INCREMENTAL FEW-SHOT LEARNING

    2023

  6. [14]

    Kukleva, A.; Kuehne, H.; and Schiele, B. 2021. Generalized and incremental few-shot learning by explicit learning and calibration without forgetting. In Proceedings of the IEEE/CVF international conference on computer vision, 9020--9029

    2021

  7. [15]

    Liu, H.; Gu, L.; Chi, Z.; Wang, Y.; Yu, Y.; Chen, J.; and Tang, J. 2022. Few-shot class-incremental learning via entropy-regularized data-free replay. In European Conference on Computer Vision, 146--162. Springer

    2022

  8. [16]

    Ma, T.-F.; Zheng, W.-L.; and Lu, B.-L. 2023. Few- Shot Class-Incremental Learning for EEG-Based Emotion Recognition . In Tanveer, M.; Agarwal, S.; Ozawa, S.; Ekbal, A.; and Jatowt, A., eds., Neural Information Processing , volume 1792, 445--455

    2023

  9. [17]

    Miotto, R.; Wang, F.; Wang, S.; Jiang, X.; and Dudley, J. T. 2018. Deep learning for healthcare: review, opportunities and challenges. Briefings in bioinformatics, 19(6): 1236--1246

    2018

  10. [18]

    Mordvintsev, A.; Olah, C.; and Tyka, M. 2015. Inceptionism: Going deeper into neural networks. Google research blog, 20(14): 5

    2015

  11. [19]

    M.; Boylan, G

    Murray, D. M.; Boylan, G. B.; Ryan, C. A.; and Connolly, S. 2009. Early EEG findings in hypoxic-ischemic encephalopathy predict outcomes at 2 years. Pediatrics, 124(3): e459--e467

    2009

  12. [20]

    M.; Mathieson, S

    O’toole, J. M.; Mathieson, S. R.; Raurale, S. A.; Magarelli, F.; Marnane, W. P.; Lightbody, G.; and Boylan, G. B. 2023. Neonatal EEG graded for severity of background abnormalities in hypoxic-ischaemic encephalopathy. Scientific Data, 10(1): 129

    2023

  13. [21]

    Peng, C.; Zhao, K.; Wang, T.; Li, M.; and Lovell, B. C. 2022. Few-shot class-incremental learning from an open-set perspective. In European Conference on Computer Vision, 382--397. Springer

    2022

  14. [22]

    Pradhan, A.; He, J.; and Jiang, N. 2022. Multi-day dataset of forearm and wrist electromyogram for hand gesture recognition and biometrics. Scientific data, 9(1): 733

    2022

  15. [23]

    H.; Suganthan, P

    Qiao, Z.; Pham, Q.; Cao, Z.; Le, H. H.; Suganthan, P. N.; Jiang, X.; and Savitha, R. 2024. Class-incremental learning for time series: Benchmark and evaluation. arXiv preprint arXiv:2402.12035

    2024 arXiv

  16. [24]

    Ramanujam, E.; Perumal, T.; and Padmavathi, S. 2021. Human activity recognition with smartphone and wearable sensors using deep learning techniques: A review. IEEE Sensors Journal, 21(12): 13029--13040

    2021

  17. [25]

    Rav \` , D.; Wong, C.; Deligianni, F.; Berthelot, M.; Andreu-Perez, J.; Lo, B.; and Yang, G.-Z. 2016. Deep learning for health informatics. IEEE journal of biomedical and health informatics, 21(1): 4--21

    2016

  18. [26]

    R.; and Yamauchi, K

    Shankarampeta, A. R.; and Yamauchi, K. 2021. Few-Shot Class Incremental Learning with Generative Feature Replay. In ICPRAM, 259--267

    2021

  19. [27]

    Shi, G.; Chen, J.; Zhang, W.; Zhan, L.-M.; and Wu, X.-M. 2021. Overcoming catastrophic forgetting in incremental few-shot learning by finding flat minima. Advances in neural information processing systems, 34: 6747--6761

    2021

  20. [28]

    Snell, J.; Swersky, K.; and Zemel, R. 2017. Prototypical networks for few-shot learning. Advances in neural information processing systems, 30

    2017

  21. [29]

    Song, Y.; Zheng, Q.; Liu, B.; and Gao, X. 2023 a . EEG Conformer : Convolutional Transformer for EEG Decoding and Visualization . IEEE Transactions on Neural Systems and Rehabilitation Engineering, 31: 710--719

    2023

  22. [30]

    Song, Z.; Zhao, Y.; Shi, Y.; Peng, P.; Yuan, L.; and Tian, Y. 2023 b . Learning with Fantasy : Semantic-Aware Virtual Contrastive Constraint for Few-Shot Class-Incremental Learning . In 2023 IEEE / CVF Conference on Computer Vision and Pattern Recognition ( CVPR ) , 24183--24192

    2023

  23. [31]

    Sun, C.; Shrivastava, A.; Singh, S.; and Gupta, A. 2017. Revisiting unreasonable effectiveness of data in deep learning era. In Proceedings of the IEEE international conference on computer vision, 843--852

    2017

  24. [32]

    Sun, L.; Zhang, M.; Wang, B.; and Tiwari, P. 2023. Few-Shot Class-Incremental Learning for Medical Time Series Classification. IEEE Journal of Biomedical and Health Informatics

    2023

  25. [33]

    Tao, X.; Hong, X.; Chang, X.; Dong, S.; Wei, X.; and Gong, Y. 2020. Few- Shot Class-Incremental Learning . In Proceedings of the IEEE / CVF Conference on Computer Vision and Pattern Recognition , 12183--12192

    2020

  26. [34]

    Van der Maaten, L.; and Hinton, G. 2008. Visualizing data using t-SNE. Journal of machine learning research, 9(11)

    2008

  27. [35]

    Wang, Q.-W.; Zhou, D.-W.; Zhang, Y.-K.; Zhan, D.-C.; and Ye, H.-J. 2024. Few-shot class-incremental learning via training-free prototype calibration. Advances in Neural Information Processing Systems, 36

    2024

  28. [36]

    Wang, X.; Wang, X.; Jiang, B.; and Luo, B. 2023. Few- Shot Learning Meets Transformer : Unified Query-Support Transformers for Few-Shot Classification . IEEE Transactions on Circuits and Systems for Video Technology

    2023

  29. [37]

    Yang, Y.; Yuan, H.; Li, X.; Lin, Z.; Torr, P.; and Tao, D. 2023. NEURAL COLLAPSE INSPIRED FEATURE-CLASSIFIER ALIGNMENT FOR FEW-SHOT CLASS INCREMENTAL LEARNING

    2023

  30. [38]

    Yao, G.; Zhu, J.; Zhou, W.; and Li, J. 2022. Few-shot class-incremental learning based on representation enhancement. Journal of Electronic Imaging, 31(4): 043027--043027

    2022

  31. [39]

    M.; Mallya, A.; Hoiem, D.; Jha, N

    Yin, H.; Molchanov, P.; Li, Z.; Alvarez, J. M.; Mallya, A.; Hoiem, D.; Jha, N. K.; and Kautz, J. 2020. Dreaming to Distill : Data-free Knowledge Transfer via DeepInversion

    2020

  32. [40]

    Zhang, S.; Li, Y.; Zhang, S.; Shahabi, F.; Xia, S.; Deng, Y.; and Alshurafa, N. 2022. Deep learning in human activity recognition with wearable sensors: A review on advances. Sensors, 22(4): 1476

    2022

  33. [41]

    Zhang, X.; Yao, L.; Wang, X.; Monaghan, J.; Mcalpine, D.; and Zhang, Y. 2019. A survey on deep learning based brain computer interface: Recent advances and new frontiers. arXiv preprint arXiv:1905.04149, 66

    2019 arXiv

  34. [42]

    Zhao, L.; Lu, J.; Xu, Y.; Cheng, Z.; Guo, D.; Niu, Y.; and Fang, X. 2023. Few-shot class-incremental learning via class-aware bilateral distillation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 11838--11847

    2023

  35. [43]

    Zhou, D.-W.; Wang, F.-Y.; Ye, H.-J.; Ma, L.; Pu, S.; and Zhan, D.-C. 2022 a . Forward compatible few-shot class-incremental learning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 9046--9056

    2022

  36. [44]

    Zhou, D.-W.; Ye, H.-J.; Ma, L.; Xie, D.; Pu, S.; and Zhan, D.-C. 2022 b . Few-shot class-incremental learning by sampling multi-phase tasks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(11): 12816--12831

    2022

  37. [45]

    Zhu, J.; Yao, G.; Zhou, W.; Zhang, G.; Ping, W.; and Zhang, W. 2022. Feature distribution distillation-based few shot class incremental learning. In 2022 5th International Conference on Pattern Recognition and Artificial Intelligence (PRAI), 108--113. IEEE

    2022

  38. [46]

    Zhuang, H.; Weng, Z.; He, R.; Lin, Z.; and Zeng, Z. 2023. GKEAL : Gaussian Kernel Embedded Analytic Learning for Few-Shot Class Incremental Task . In 2023 IEEE / CVF Conference on Computer Vision and Pattern Recognition ( CVPR ) , 7746--7755

    2023

  39. [47]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all...

  40. [48]

    write newline

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

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.