Pith. sign in

REVIEW 5 major objections 6 minor 44 references

Maintain Plasticity in Long-timescale Continual Test-time Adaptation

T0 review · 5 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Long continual test-time adaptation collapses as models lose plasticity, and label-flip fluctuations mark the collapse; a shrink-restore reset timed to them reports 40.0 average accuracy versus 37.9.

desk verdict A pragmatic adaptive-reset method with real benchmark gains, but the trigger is under-specified and the label-flip signal is backed by thin evidence. read the letter →

arxiv 2412.20034 v2 pith:VYLUQI2M submitted 2024-12-28 cs.CV

Yanshuo Wang , Xuesong Li , Jinguang Tong , Jie Hong , Jun Lan , Weiqiang Wang , Huijia Zhu , Haoxing Chen This is my paper · ORCID
classification cs.CV
keywords continualtest-timeadaptationplasticitylabelflipadaptivere-initializationshrink-restorenon-stationarydomainshiftentropyminimization
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

Continual test-time adaptation (CTTA) aims to keep a pre-trained model accurate as the test stream keeps changing, but the paper shows that over very long streams most existing methods steadily lose plasticity, the ability to keep learning from new data. The paper's claim is that this loss is telegraphed by the label flip—the count of images whose predicted class changes between consecutive model states—whose smoothed trajectory starts to fluctuate just as accuracy bottoms out. To act on that signal, the paper proposes Adaptive Shrink-Restore (ASR): when the smoothed label flip rises above a multiple $\pi$ of its running minimum, it reinitializes the weights as $\theta_{\mathrm{reinit}} = \lambda\theta_t + \gamma\theta_{\mathrm{source}}$ with $\lambda=0.2$ and $\gamma=0.75$, shrinking what the model has learned and restoring a share of the source weights. On the CIN-C, CIN-3DCC, and CCC benchmarks this schedule reports an average accuracy of 40.0, ahead of the 37.9 of the fixed-interval reset baseline RDumb, without needing a validation set to tune the reset interval.

What carries the argument

The load-bearing mechanism is the label-flip signal paired with a threshold trigger, plus the shrink-restore update that follows. The label flip at step $t$ is $LF_t = \sum_i I(i)\,c_i\,(c_{i,t}-c_{i,t-1})$, where $I(i)$ indicates a predicted-class change for sample $i$ and the $c$'s are confidence values from the current and previous model; the raw signal is smoothed by an exponential moving average with $\beta=0.8$. The trigger computes $Min$ as the mean of the smoothed flip around its lowest point and declares a reset when $LF_{t+1} > \pi\,Min$ for a threshold $\pi$; this is what turns the observation into a schedule that needs no validation set. The reset then applies $\theta_{\mathrm{reinit}} = \lambda\theta_t + \gamma\theta_{\mathrm{source}}$ with $\lambda=0.2$, $\gamma=0.75$, and $\lambda+\gamma<1$, shrinking the adapted weights, restoring a scaled source initialization, and bounding weight magnitude over time. The whole chain converts a measurable, model-intrinsic quantity—how often predictions flip—into a decision about when to trade accumulated knowledge for fresh learning capacity.

What would settle it

On a long CCC-medium stream, record at each step the smoothed label flip $LF_t$, its running minimum $Min$, and an independent plasticity probe, e.g., how quickly the model can fit a held-out batch of the current domain in a few gradient steps. If for a range of thresholds $\pi$ the trigger fires while the probe still shows strong learning, or stays silent while the probe shows learning has stalled, then label-flip fluctuation is not a faithful trigger of plasticity loss.

Watch

Extended reading notes

Core claim

The central claim, stated on the paper's own terms, is that the long-run collapse of CTTA models is a plasticity failure rather than just a domain-mismatch failure, and that the model's label-flip trajectory can be read as a gauge of that plasticity. Using classification accuracy as the plasticity metric, the paper shows that entropy-minimizing and pseudo-labeling methods lose accuracy at an accelerating rate relative to a periodically reset model. It then identifies the trigger: after an initial decreasing phase, the smoothed label flip begins to fluctuate, and the minimum accuracy is reached around the moment the fluctuation starts; re-initialization fires when $LF_{t+1} > \pi \cdot Min$, where $Min$ is the mean of the smoothed label flip around its lowest point. The reset itself is a shrink-restore step, $\theta_{\mathrm{reinit}} = \lambda\theta_t + \gamma\theta_{\mathrm{source}}$ with $\lambda + \gamma < 1$, which keeps previous knowledge as a warm start while injecting source-model plasticity and preventing unbounded weight growth. Reported results put ASR's average accuracy at 40.0 across the three benchmarks, above RDumb's 37.9, with the largest margin on the hardest splits.

Load-bearing premise

The load-bearing premise is that a rise in the smoothed label-flip count above $\pi$ times its running minimum reliably marks the moment plasticity is lost; the paper argues for this in Section 3.2.3 from a qualitative trajectory plot and does not report a value for $\pi$ or a quantitative correlation between flips and plasticity.

Editorial extensions

If this is right

  • ASR removes the need for a separate validation set to pick the reset interval: timing comes from the model's own label-flip trajectory, and the paper shows the adaptive trigger matches a best-tuned fixed interval on Tent, EATA, and RPL.
  • Because the reset rule is wrapper-like, any CTTA method built on entropy or pseudo-label minimization can carry the ASR mechanism, so the plasticity-preservation benefit is not tied to a single optimizer.
  • The method's largest gains appear on the hardest splits (CCC-Hard: 12.9 vs 9.6 for RDumb), suggesting that adaptive re-initialization matters most when target distributions drift farthest from the source.
  • The $\lambda+\gamma<1$ constraint keeps adapted weights from growing without bound, directly addressing the weight-explosion failure mode identified for long CTTA runs.
  • If the trigger genuinely tracks plasticity, then fixed-interval resets are overkill: they disturb adaptation either too early or too late, while the adaptive schedule intervenes only when learning capacity is actually impaired.

Reading between the lines

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

  • Not tested in the paper, but consistent with its mechanism, label-flip fluctuation could serve as a general plasticity probe for other online-learning settings such as continual reinforcement learning or streaming fine-tuning, where reset timing is usually chosen heuristically.
  • A stronger quantitative test of the paper's association would compute a lagged correlation between the smoothed flip trajectory and an independent plasticity metric, such as the ability to fit a held-out probe batch; the paper currently supports the link with a visual alignment rather than a measured correlation.
  • Because the trigger threshold $\pi$ is one global value, its scale depends on the base method's confidence distribution; normalizing $\pi$ by the running distribution of flips, or triggering on a percentile jump, would likely make ASR more portable across methods and datasets.
  • The shrink-restore target is always the source checkpoint, so the method implicitly assumes source weights stay a useful plasticity reservoir even after long adaptation; if that assumption erodes on extremely long streams, a reservoir updated from recent well-performing weights might work better.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 6 minor

Summary. The paper studies plasticity in continual test-time adaptation (CTTA) and proposes Adaptive Shrink-Restore (ASR), a policy that reinitializes model weights at adaptively determined times using a label-flip trajectory. The adaptive timing is based on the observation that plasticity loss coincides with fluctuations in the label-flip count. The reinitialization step combines shading the current weights with restoring source weights. On three benchmarks (CIN-C, CIN-3DCC, and CCC), ASR reports an average accuracy of 40.0 versus 37.9 for the RDumb baseline, with larger gains on harder CCC splits. The paper also argues that ASR can be plugged into any CTTA method.

Significance. If the central claim holds, ASR provides a validation-free adaptive reset schedule, which is a practical contribution to long-timescale CTTA where fixed-interval resets require external tuning. The paper is the first to explicitly study plasticity in CTTA and gives evidence that most existing methods lose plasticity over long streams. The method is simple, general, and shows consistent improvements across several large-scale benchmarks, which are strengths. However, the load-bearing association between label-flip fluctuations and plasticity loss is supported only by a qualitative figure, the trigger hyperparameters are not reported, and the ablation of the adaptive trigger is not quantitative. These issues limit the reproducibility and the strength of the central claim, but they are addressable within the scope of a revision.

major comments (5)
  1. [Section 3.2.3, Eqs. (3)-(4), Fig. 5] The assertion that severe plasticity loss happens when the label flip starts to fluctuate is supported only by a qualitative reading of Fig. 5. No correlation statistic, trigger precision/recall, or error bars are reported. The trigger thresholds π and the neighborhood size N are never specified in Section 4.2 or elsewhere, making the method non-reproducible. Additionally, the running minimum bookkeeping is not described; as written, Eq. (4) could re-fire immediately after a reinitialization unless the minimum or smoothed state is reset, which is not stated.
  2. [Section 3.2.2, Eq. (2)] The exponential moving average in Eq. (2) has an indexing inconsistency: LF_{t+1} = β LF_{t-1} + (1−β) LF_t mixes time indices and does not define a standard recursion. The authors should rewrite the EMA with a clear smoothed series, for example S_{t+1} = β S_t + (1−β) LF_{t+1}, and apply Eqs. (3)-(4) consistently to that smoothed series. The current ambiguity affects the definition of the trigger itself.
  3. [Section 4.4.1, Fig. 3, Table 3] The ablation of the adaptive trigger is qualitative. The claim that the adaptive trigger alone performs similarly to the best fixed interval is not supported by a table with numbers; the 'Adp' column in Table 3 is not described, and it is unclear what adaptive trigger was used and with which hyperparameters. Please provide a quantitative comparison of the trigger alone (without shrink-restore) against fixed reset intervals, including the best fixed interval and a random-timing baseline, with standard deviations across the CCC combinations.
  4. [Table 2, Section 4.4.3] The shrink-restore ablation is conducted on a 902,200-image subset of CCC-Medium and shows an improvement of only 0.48 accuracy points, while the main result in Table 1 reports a 2.9-point average improvement over RDumb. Without a full ablation matrix (base, +trigger, +shrink-restore, +both) on the complete CCC splits, the independent contribution of each component cannot be assessed, and the source of the reported improvement remains unclear.
  5. [Section 3.1, Section 3.2.3] The paper measures plasticity by classification accuracy and then uses accuracy curves to claim an association between label-flip fluctuations and plasticity loss. Because both the trigger signal and the evaluation metric derive from the model's own predictions, the association is at least partly circular. To support the causal claim, the authors should validate the trigger against an independent measure of learning capacity (e.g., the accuracy of a freshly reset probe or the ability to fit a new batch) or report a quantitative time-series correlation between label-flip fluctuation and accuracy decline.
minor comments (6)
  1. [Section 4.2] The values of π and N for the adaptive trigger are missing; they should be listed together with β, λ, and γ in the implementation paragraph.
  2. [Section 3.2.2] The label-flip computation says 'we run a set of test images' without specifying which images, how many, and how often this is computed; please clarify whether it is the current batch or a separate reference set.
  3. [Table 3 caption] The caption refers to 'various tent methods', but the rows include TENT, ETA, and RPL; please correct the caption to match the actual rows.
  4. [Fig. 5 and surrounding text] The purple line and blue line are not consistently described in the text and caption; please clearly specify which curve corresponds to label flip and which to accuracy in every mention.
  5. [Section 2.2] There is a typo: 'serval methods' should be 'several methods'.
  6. [Eq. (1)] The summation in Eq. (1) does not explicitly state whether it is over images in a batch or over a fixed set; please clarify the index range.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity; ASR's components are independently defined and benchmarked, with only a minor non-load-bearing self-citation.

full rationale

The paper's central claim is that the ASR policy, combining an adaptive label-flip trigger and a shrink-restore weight update, improves long-timescale CTTA performance. This claim is evaluated on external benchmarks (CIN-C, CIN-3DCC, CCC) against standard baselines, so the headline result is self-contained rather than derived from the paper's own definitions. The shrink-restore update in Eq. (5) is an independent mechanism, motivated by prior work on 'shrink and perturb' but not derived from the label-flip signal. The adaptive trigger in Eqs. (1)-(4) is a heuristic change-point detector; the asserted correlation between label-flip fluctuations and plasticity loss is presented as an empirical observation (Fig. 5), not as an identity. Plasticity is operationalized as classification accuracy (Sec. 3.1), while label flip is a prediction-change statistic, so there is no equation-level equivalence that forces the claimed association. The only self-citation is reference [41], the first author's prior CTTA work, cited as an example of existing methods; it is not load-bearing for the method or results, and it appears to be a citation error since CoTTA is reference [40]. The paper does leave the trigger parameters pi and N unreported, which is a reproducibility weakness, and the qualitative validation of the trigger is thin, but these are correctness concerns, not circularity. Overall, the derivation chain does not reduce any prediction to its inputs by construction.

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

The method introduces no new physical entities. The free parameters are hand-set or unreported, and the central trigger rests on the unproven assumption that label flip is a reliable plasticity-loss signal.

free parameters (5)
  • beta (EMA coefficient) = 0.8
    Hand-set smoothing factor for the label-flip moving average in Eq. 2; no ablation reported.
  • lambda (shrink coefficient) = 0.2
    Hand-set scale applied to current weights in the shrink-restore update, Eq. 5.
  • gamma (restore coefficient) = 0.75
    Hand-set scale applied to source weights in the shrink-restore update, Eq. 5.
  • pi (fluctuation threshold) = not reported
    The trigger fires when LF > pi * Min in Eq. 4; the value is never given, under-specifying the method.
  • N (neighborhood size for minimum) = not reported
    The number of points around the lowest label-flip value used in Eq. 3 is not specified.
assumptions (3)
  • domain assumption Classification accuracy is a valid proxy for model plasticity (following Wang et al. 2024, ref [39]).
    Section 3.1 states 'we use the classification accuracy to represent the model plasticity.' This equates the learning-capacity concept with a single evaluation metric.
  • ad hoc to paper Label-flip fluctuations signal plasticity loss.
    Section 3.2.3 asserts 'severe plasticity loss happens when the label flip meets fluctuations' based on one qualitative trajectory in Fig. 5; no statistical test is provided.
  • ad hoc to paper Source model weights are a good complement to preserve plasticity because domains are correlated.
    Section 3.2.4 justifies restoring the source weights with this argument; the benefit is only empirically suggested by a single ablation on a subset.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Maintain Plasticity in Long-timescale Continual Test-time Adaptation." pith.science (2026). https://pith.science/paper/VYLUQI2M

@misc{pith2026241220034,
  author       = {Pith},
  title        = {Pith review of: Maintain Plasticity in Long-timescale Continual Test-time Adaptation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VYLUQI2M}},
  note         = {Machine review of arXiv:2412.20034}
}
read the original abstract

Continual test-time domain adaptation (CTTA) aims to adjust pre-trained source models to perform well over time across non-stationary target environments. While previous methods have made considerable efforts to optimize the adaptation process, a crucial question remains: can the model adapt to continually-changing environments with preserved plasticity over a long time? The plasticity refers to the model's capability to adjust predictions in response to non-stationary environments continually. In this work, we explore plasticity, this essential but often overlooked aspect of continual adaptation to facilitate more sustained adaptation in the long run. First, we observe that most CTTA methods experience a steady and consistent decline in plasticity during the long-timescale continual adaptation phase. Moreover, we find that the loss of plasticity is strongly associated with the change in label flip. Based on this correlation, we propose a simple yet effective policy, Adaptive Shrink-Restore (ASR), towards preserving the model's plasticity. In particular, ASR does the weight re-initialization by the adaptive intervals. The adaptive interval is determined based on the change in label flipping. Our method is validated on extensive CTTA benchmarks, achieving excellent performance.

Figures

Figures reproduced from arXiv: 2412.20034 by the authors.

Figure 1
Figure 1. Illustration of plasticity loss in adaptation: Test data continuously changes from one domain to another, and the dotted line [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. A simple test of different TTA methods on the part of the CCC-medium dataset for plasticity loss. We observe that all tested [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Model plasticity under different plasticity preservation policy timings: Using RPL [ [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: The framework of our proposed ASR: when continuously adapting in non-stationary domains, reinitialization is performed [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: The proposed adaptive trigger policy is used to determine the suitable point for initialization. When significant fluctuations in [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 29 canonical work pages

  1. [1]

    Synaptic plasticity: taming the beast

    Larry F Abbott and Sacha B Nelson. Synaptic plasticity: taming the beast. Nature neuroscience, 3(11):1178–1183,

    work page

  2. [2]

    On warm-starting neural network training

    Jordan Ash and Ryan P Adams. On warm-starting neural network training. Advances in neural information processing systems, 33:3884–3894, 2020. 3, 5

    work page 2020

  3. [3]

    The influence of age of acquisition in word reading and other tasks: A never ending story? Journal of Memory and language, 50(4):456–476, 2004

    Patrick Bonin, Christopher Barry, Alain M ´eot, and Maryl`ene Chalard. The influence of age of acquisition in word reading and other tasks: A never ending story? Journal of Memory and language, 50(4):456–476, 2004. 3

    work page 2004

  4. [4]

    Autodial: Automatic domain alignment layers

    Fabio Maria Carlucci, Lorenzo Porzi, Barbara Caputo, Elisa Ricci, and Samuel Rota Bulo. Autodial: Automatic domain alignment layers. In 2017 IEEE international conference on computer vision (ICCV), pages 5077–5085. IEEE, 2017. 2

    work page 2017

  5. [5]

    Imagenet: A large-scale hierarchical image database

    Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition, pages 248–255. Ieee, 2009. 6

    2009

  6. [6]

    Re- ducing flipping errors in deep neural networks

    Xiang Deng, Yun Xiao, Bo Long, and Zhongfei Zhang. Re- ducing flipping errors in deep neural networks. In Proceed- ings of the AAAI Conference on Artificial Intelligence , vol- ume 36, pages 6506–6514, 2022. 3

    work page 2022

  7. [7]

    Robust mean teacher for continual and gradual test-time adaptation

    Mario D ¨obler, Robert A Marsden, and Bin Yang. Robust mean teacher for continual and gradual test-time adaptation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 7704–7714, 2023. 1, 2

    work page 2023

  8. [8]

    Loss of plasticity in deep continual learning

    Shibhansh Dohare, J Fernando Hernandez-Garcia, Qingfeng Lan, Parash Rahman, A Rupam Mahmood, and Richard S Sutton. Loss of plasticity in deep continual learning. Nature, 632(8026):768–774, 2024. 1, 3

    work page 2024

Show all 44 references
  1. [9]

    Domain-adversarial training of neural networks

    Yaroslav Ganin, Evgeniya Ustinova, Hana Ajakan, Pas- cal Germain, Hugo Larochelle, Franc ¸ois Laviolette, Mario March, and Victor Lempitsky. Domain-adversarial training of neural networks. Journal of machine learning research , 17(59):1–35, 2016. 2

    2016

  2. [10]

    Test time adaptation via conjugate pseudo-labels.Ad- vances in Neural Information Processing Systems, 35:6204– 6218, 2022

    Sachin Goyal, Mingjie Sun, Aditi Raghunathan, and J Zico Kolter. Test time adaptation via conjugate pseudo-labels.Ad- vances in Neural Information Processing Systems, 35:6204– 6218, 2022. 3, 6, 7

    2022

  3. [11]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016. 6

    2016

  4. [12]

    Benchmarking neu- ral network robustness to common corruptions and perturba- tions

    Dan Hendrycks and Thomas Dietterich. Benchmarking neu- ral network robustness to common corruptions and perturba- tions. arXiv preprint arXiv:1903.12261, 2019. 3, 6

    1903 arXiv

  5. [13]

    3d common corruptions and data augmentation

    O ˘guzhan Fatih Kar, Teresa Yeo, Andrei Atanov, and Amir Zamir. 3d common corruptions and data augmentation. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 18963–18974, 2022. 6

    2022

  6. [14]

    Implicit under-parameterization inhibits data- efficient deep reinforcement learning

    Aviral Kumar, Rishabh Agarwal, Dibya Ghosh, and Sergey Levine. Implicit under-parameterization inhibits data- efficient deep reinforcement learning. arXiv preprint arXiv:2010.14498, 2020. 3

    2010 arXiv

  7. [15]

    Adaptive batch normalization for practical do- main adaptation

    Yanghao Li, Naiyan Wang, Jianping Shi, Xiaodi Hou, and Jiaying Liu. Adaptive batch normalization for practical do- main adaptation. Pattern Recognition, 80:109–117, 2018. 2

    2018

  8. [16]

    Consistent123: One image to highly consistent 3d asset using case-aware diffusion priors

    Yukang Lin, Haonan Han, Chaoqun Gong, Zunnan Xu, Yachao Zhang, and Xiu Li. Consistent123: One image to highly consistent 3d asset using case-aware diffusion priors. In Proceedings of the 32nd ACM International Conference on Multimedia, pages 6715–6724, 2024. 1

    2024

  9. [17]

    Learning transferable features with deep adaptation net- works

    Mingsheng Long, Yue Cao, Jianmin Wang, and Michael Jor- dan. Learning transferable features with deep adaptation net- works. In International conference on machine learning , pages 97–105. PMLR, 2015. 1

    2015

  10. [18]

    Conditional adversarial domain adapta- tion

    Mingsheng Long, Zhangjie Cao, Jianmin Wang, and Michael I Jordan. Conditional adversarial domain adapta- tion. Advances in neural information processing systems , 31, 2018. 2

    2018

  11. [19]

    Switching between tasks can cause ai to lose the ability to learn, 2024

    Clare Lyle and Razvan Pascanu. Switching between tasks can cause ai to lose the ability to learn, 2024. 1, 3

    2024

  12. [20]

    Understanding plasticity in neural networks

    Clare Lyle, Zeyu Zheng, Evgenii Nikishin, Bernardo Avila Pires, Razvan Pascanu, and Will Dabney. Understanding plasticity in neural networks. In International Conference on Machine Learning, pages 23190–23211. PMLR, 2023. 1, 3

    2023

  13. [21]

    The stability-plasticity dilemma: Investigating the contin- uum from catastrophic forgetting to age-limited learning ef- fects, 2013

    Martial Mermillod, Aur ´elia Bugaiska, and Patrick Bonin. The stability-plasticity dilemma: Investigating the contin- uum from catastrophic forgetting to age-limited learning ef- fects, 2013. 2

    2013

  14. [22]

    Test-time adaptation to distribution shift by confi- dence maximization and input transformation.arXiv preprint arXiv:2106.14999, 2021

    Chaithanya Kumar Mummadi, Robin Hutmacher, Kilian Rambach, Evgeny Levinkov, Thomas Brox, and Jan Hendrik Metzen. Test-time adaptation to distribution shift by confi- dence maximization and input transformation.arXiv preprint arXiv:2106.14999, 2021. 6, 7

    2021 arXiv

  15. [23]

    Evaluating prediction-time batch normalization for robust- ness under covariate shift

    Zachary Nado, Shreyas Padhy, D Sculley, Alexander D’Amour, Balaji Lakshminarayanan, and Jasper Snoek. Evaluating prediction-time batch normalization for robust- ness under covariate shift. arXiv preprint arXiv:2006.10963,

    2006 arXiv

  16. [24]

    Deep re- inforcement learning with plasticity injection

    Evgenii Nikishin, Junhyuk Oh, Georg Ostrovski, Clare Lyle, Razvan Pascanu, Will Dabney, and Andr´e Barreto. Deep re- inforcement learning with plasticity injection. Advances in Neural Information Processing Systems, 36, 2024. 3

    2024

  17. [25]

    Efficient test-time model adaptation without forgetting

    Shuaicheng Niu, Jiaxiang Wu, Yifan Zhang, Yaofo Chen, Shijian Zheng, Peilin Zhao, and Mingkui Tan. Efficient test-time model adaptation without forgetting. In Interna- tional conference on machine learning, pages 16888–16905. PMLR, 2022. 1, 2, 3, 6, 7, 8

    2022

  18. [26]

    Rdumb: A simple approach that questions our progress in continual test-time adaptation

    Ori Press, Steffen Schneider, Matthias K ¨ummerer, and Matthias Bethge. Rdumb: A simple approach that questions our progress in continual test-time adaptation. Advances in Neural Information Processing Systems, 36, 2024. 1, 3, 6, 7, 8

    2024

  19. [27]

    The entropy enigma: Success and failure of entropy minimization

    Ori Press, Ravid Shwartz-Ziv, Yann LeCun, and Matthias Bethge. The entropy enigma: Success and failure of entropy minimization. arXiv preprint arXiv:2405.05012, 2024. 1, 3

    2024 arXiv

  20. [28]

    Learning to learn without forgetting by maximizing transfer and minimizing interference

    Matthew Riemer, Ignacio Cases, Robert Ajemian, Miao Liu, Irina Rish, Yuhai Tu, and Gerald Tesauro. Learning to learn without forgetting by maximizing transfer and minimizing interference. arXiv preprint arXiv:1810.11910, 2018. 3

    arXiv

  21. [29]

    If your data distribution shifts, use self- learning

    Evgenia Rusak, Steffen Schneider, George Pachitariu, Luisa Eck, Peter Gehler, Oliver Bringmann, Wieland Brendel, and Matthias Bethge. If your data distribution shifts, use self- learning. arXiv preprint arXiv:2104.12928, 2021. 4, 6, 7

    2021 arXiv

  22. [30]

    Improving robustness against common corruptions by covariate shift adaptation

    Steffen Schneider, Evgenia Rusak, Luisa Eck, Oliver Bring- mann, Wieland Brendel, and Matthias Bethge. Improving robustness against common corruptions by covariate shift adaptation. Advances in neural information processing sys- tems, 33:11539–11551, 2020. 2, 6, 7

    2020

  23. [31]

    The dormant neuron phenomenon in deep rein- forcement learning

    Ghada Sokar, Rishabh Agarwal, Pablo Samuel Castro, and Utku Evci. The dormant neuron phenomenon in deep rein- forcement learning. In International Conference on Machine Learning, pages 32145–32168. PMLR, 2023. 3

    2023

  24. [32]

    Ecotta: Memory-efficient continual test-time adaptation via self-distilled regularization

    Junha Song, Jungsoo Lee, In So Kweon, and Sungha Choi. Ecotta: Memory-efficient continual test-time adaptation via self-distilled regularization. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 11920–11929, 2023. 1

    2023

  25. [33]

    Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results

    Antti Tarvainen and Harri Valpola. Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results. Advances in neural information processing systems, 30, 2017. 2

    2017

  26. [34]

    An empirical study of example forget- ting during deep neural network learning

    Mariya Toneva, Alessandro Sordoni, Remi Tachet des Combes, Adam Trischler, Yoshua Bengio, and Geof- frey J Gordon. An empirical study of example forget- ting during deep neural network learning. arXiv preprint arXiv:1812.05159, 2018. 3

    2018 arXiv

  27. [35]

    Learning to adapt structured output space for semantic seg- mentation

    Yi-Hsuan Tsai, Wei-Chih Hung, Samuel Schulter, Ki- hyuk Sohn, Ming-Hsuan Yang, and Manmohan Chandraker. Learning to adapt structured output space for semantic seg- mentation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 7472–7481,

  28. [36]

    Adversarial discriminative domain adaptation

    Eric Tzeng, Judy Hoffman, Kate Saenko, and Trevor Darrell. Adversarial discriminative domain adaptation. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 7167–7176, 2017. 2

    2017

  29. [37]

    Tent: Fully test-time adaptation by entropy minimization

    Dequan Wang, Evan Shelhamer, Shaoteng Liu, Bruno Ol- shausen, and Trevor Darrell. Tent: Fully test-time adaptation by entropy minimization. arXiv preprint arXiv:2006.10726,

    2006 arXiv

  30. [38]

    A comprehensive survey of continual learning: theory, method and application

    Liyuan Wang, Xingxing Zhang, Hang Su, and Jun Zhu. A comprehensive survey of continual learning: theory, method and application. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024. 3

    2024

  31. [39]

    Improving plasticity in online continual learning via collaborative learning

    Maorong Wang, Nicolas Michel, Ling Xiao, and Toshihiko Yamasaki. Improving plasticity in online continual learning via collaborative learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 23460–23469, 2024. 3

    2024

  32. [40]

    Continual test-time domain adaptation

    Qin Wang, Olga Fink, Luc Van Gool, and Dengxin Dai. Continual test-time domain adaptation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 7201–7211, 2022. 2, 3, 6, 7

    2022

  33. [41]

    Continual test-time domain adaptation via dynamic sample selection

    Yanshuo Wang, Jie Hong, Ali Cheraghian, Shafin Rahman, David Ahmedt-Aristizabal, Lars Petersson, and Mehrtash Harandi. Continual test-time domain adaptation via dynamic sample selection. In Proceedings of the IEEE/CVF Win- ter Conference on Applications of Computer Vision , pa...

    2024

  34. [42]

    Age of acquisition effects in word reading and other tasks

    Jason D Zevin and Mark S Seidenberg. Age of acquisition effects in word reading and other tasks. Journal of Memory and language, 47(1):1–29, 2002. 2, 3

    2002

  35. [43]

    Unsupervised domain adaptation for semantic segmentation via class-balanced self-training

    Yang Zou, Zhiding Yu, BVK Kumar, and Jinsong Wang. Unsupervised domain adaptation for semantic segmentation via class-balanced self-training. In Proceedings of the Eu- ropean conference on computer vision (ECCV) , pages 289– 305, 2018. 2

    2018

  36. [44]

    Confidence regularized self-training

    Yang Zou, Zhiding Yu, Xiaofeng Liu, BVK Kumar, and Jin- song Wang. Confidence regularized self-training. In Pro- ceedings of the IEEE/CVF international conference on com- puter vision, pages 5982–5991, 2019. 2

    2019

Pith tools

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