arxiv: 2605.11617 · v1 · submitted 2026-05-12 · 💻 cs.LG · math.ST· stat.TH

Recognition: 2 theorem links

· Lean Theorem

MIST: Reliable Streaming Decision Trees for Online Class-Incremental Learning via McDiarmid Bound

Chi-Nguyen Tran, Dao Sy Duy Minh, Huynh Trung Kiet, Long Tran-Thanh, Nguyen Lam Phu Quy, Phu-Hoa Pham

Pith reviewed 2026-05-13 01:15 UTC · model grok-4.3

classification 💻 cs.LG math.STstat.TH

keywords streaming decision treesclass-incremental learningMcDiarmid boundGini impurityonline continual learningquantile sketchesBayesian inheritancenon-stationary streams

0 comments

The pith

Streaming decision trees achieve reliable splits for online class-incremental learning by using a K-independent McDiarmid bound on Gini impurity together with Bayesian inheritance and quantile sketches.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Standard streaming decision trees fail in class-incremental settings because their split criteria grow unreliable as the number of classes increases. The root problem is that Information Gain scales with the log of the class count, so any Hoeffding-style confidence interval derived from it must widen accordingly. MIST replaces this scaling with a McDiarmid bound on Gini impurity that stays independent of class count, adds a protocol to project parent node statistics onto child nodes using truncated-Gaussian moments, and stores per-leaf quantile sketches that support both threshold checks and geometry-aware predictions. If these hold, decision trees could serve as a memory-bounded, locally updating alternative to parametric continual learners, especially on non-Gaussian data where global methods often degrade. The three pieces are integrated so that conservative splits receive the strongest statistical inheritance.

Core claim

MIST resolves both the unreliability of splits and the absence of knowledge transfer in streaming decision trees for online class-incremental learning through three integrated components: a tight K-independent McDiarmid confidence radius for Gini splitting that acts as a structural regulariser, a Bayesian inheritance protocol that projects parent statistics to child nodes via truncated-Gaussian moments with variance reduction strongest precisely when splitting is most conservative, and per-leaf KLL quantile sketches that support both continuous threshold evaluation and geometry-adaptive leaf prediction from a single data structure. This combination removes the log K scaling problem inherent,

What carries the argument

The K-independent McDiarmid confidence radius applied to Gini impurity, which functions as a structural regulariser to keep split decisions reliable regardless of how many classes have appeared so far.

Load-bearing premise

The McDiarmid bound applied to Gini impurity yields a tight radius that remains independent of class count under streaming non-stationary conditions, and the truncated-Gaussian projection accurately captures parent-to-child statistics without bias that grows with class count.

What would settle it

A controlled stream in which new classes are introduced sequentially and the fraction of incorrect splits or the drop in leaf accuracy increases with class count instead of staying stable.

Figures

Figures reproduced from arXiv: 2605.11617 by Chi-Nguyen Tran, Dao Sy Duy Minh, Huynh Trung Kiet, Long Tran-Thanh, Nguyen Lam Phu Quy, Phu-Hoa Pham.

**Figure 2.** Figure 2: Tree-dynamics diagnostics on Synth-50, Covertype, and Split-MNIST (standard, un [PITH_FULL_IMAGE:figures/full_fig_p021_2.png] view at source ↗

**Figure 3.** Figure 3: 2D PCA visualisations of the eight stress-test streams (50 samples per class shown). Each [PITH_FULL_IMAGE:figures/full_fig_p027_3.png] view at source ↗

read the original abstract

Streaming decision trees are natural candidates for open-world continual learning, as they perform local updates, enjoy bounded memory, and static decision boundaries. Despite these, they still fail in online class-incremental learning due to two coupled miscalibrations: (i) their split criterion grows unreliable as the class count K expands, and (ii) the absence of knowledge transfer at split time. Both failures share a common root: the range of Information Gain intrinsically scales with log2 K. Consequently, any Hoeffding-style confidence radius derived from it must inevitably grow with the class count, making a K-independent split criterion structurally impossible, taking away the potential benefits of applying streaming decision trees to continual learning. To fix this issue, we present MIST (McDiarmid Incremental Streaming Tree), which resolves both failures through three integrated components: (i) a tight, K-independent McDiarmid confidence radius for Gini splitting that acts as a structural regulariser; (ii) a Bayesian inheritance protocol that projects parent statistics to child nodes via truncated-Gaussian moments, with variance reduction guarantees strongest precisely when splitting is most conservative; and (iii) per-leaf KLL quantile sketches that support both continuous threshold evaluation and geometry-adaptive leaf prediction from a single data structure. On standard and stress-test tabular streams, MIST is competitive with global parametric methods on near-Gaussian benchmarks and uniquely robust on non-Gaussian geometry where SOTA benchmarks collapse.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

MIST has an interesting approach to K-independent Gini bounds for streaming trees in continual learning, but the independence claim needs verification against non-stationary class growth.

read the letter

The key takeaway is that this paper tries to make streaming decision trees work for online class-incremental learning by deriving a McDiarmid bound on the Gini split criterion that stays independent of the class count K. They combine it with Bayesian inheritance via truncated Gaussians and KLL sketches to handle knowledge transfer and leaf predictions. This is new in the way it targets the root cause of the log K scaling in information gain bounds. The approach gives a structural regularizer through the bound and claims stronger variance reduction when splits are conservative. On the positive side, it shows results competitive with global models on Gaussian-like data and better robustness on non-Gaussian cases, which is a practical win if it holds. The soft spot is the validity of that K-independent radius under the paper's own conditions. As K grows with new classes and the stream is non-stationary, the bounded difference constants in McDiarmid for Gini could still depend on K through the probability support. If the derivation doesn't explicitly handle every increment in K, the bound reintroduces the scaling problem the authors want to avoid. The truncated Gaussian projection also assumes a form that might not fit all distributions. The abstract is light on the actual math, so this needs checking. This work is aimed at the continual learning community, especially those interested in non-parametric models for tabular streams. A reader working on streaming algorithms or decision trees would get value from the proposed components and the stress tests mentioned. It deserves serious peer review because the problem is real and the proposed solution is technically grounded enough to merit detailed feedback on the derivations and experiments.

Referee Report

2 major / 2 minor

Summary. The manuscript introduces MIST, a streaming decision tree for online class-incremental learning. It identifies that Hoeffding-style bounds on Information Gain become unreliable as class count K grows because the range scales with log K. MIST addresses this via three components: a claimed tight K-independent McDiarmid confidence radius on Gini impurity that serves as structural regularizer, a Bayesian inheritance protocol projecting parent node statistics to children using truncated-Gaussian moments (with variance reduction strongest at conservative splits), and per-leaf KLL quantile sketches enabling continuous threshold evaluation and geometry-adaptive leaf prediction. Experiments on standard and stress-test tabular streams report competitiveness with global parametric methods on near-Gaussian data and superior robustness on non-Gaussian geometries.

Significance. If the McDiarmid radius derivation establishes genuine K-independence without hidden dependence on class probabilities and the truncated-Gaussian projection delivers unbiased variance reduction, the work would offer a practical non-parametric alternative for continual learning settings with expanding classes. The combination of concentration inequalities, moment-based inheritance, and quantile sketches is a coherent technical integration that could extend decision-tree applicability beyond current limitations in non-stationary streams.

major comments (2)

[§3.1] §3.1 (McDiarmid Bound for Gini): The central claim requires a McDiarmid-derived radius that is both tight and strictly K-independent to act as a structural regularizer. The bounded-difference constants c_i for the Gini function are derived from class probabilities p_j; in class-incremental streaming where K increases and the distribution is non-stationary, these constants can acquire implicit K-dependence through the support size and new-class probability mass. The manuscript must supply the explicit calculation demonstrating that this dependence cancels for arbitrary K; without it the radius grows with K exactly as the Hoeffding bounds the authors criticize, undermining both the regularizer and the downstream inheritance guarantees.
[§4.2] §4.2 (Bayesian Inheritance Protocol): The truncated-Gaussian moment projection claims variance reduction that is strongest precisely when splitting is most conservative. In the class-incremental regime, new classes enter after splits have occurred; the Gaussian assumption may then be violated, introducing bias that accumulates with K. A formal bias bound or empirical verification on streams with deliberately increasing K is needed to confirm the protocol does not degrade the claimed robustness.

minor comments (2)

[Abstract] Abstract: The assertion that MIST is 'uniquely robust' on non-Gaussian geometry should be backed by explicit quantitative deltas versus the strongest baselines in the results section rather than qualitative description.
[Notation] Notation: Ensure the definition of the McDiarmid radius (including the precise form of the bounded differences) is identical between the theoretical derivation and the pseudocode/implementation description.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback on our manuscript. The comments highlight important aspects of the McDiarmid bound derivation and the Bayesian inheritance protocol that we address below with clarifications and planned revisions.

read point-by-point responses

Referee: [§3.1] §3.1 (McDiarmid Bound for Gini): The central claim requires a McDiarmid-derived radius that is both tight and strictly K-independent to act as a structural regularizer. The bounded-difference constants c_i for the Gini function are derived from class probabilities p_j; in class-incremental streaming where K increases and the distribution is non-stationary, these constants can acquire implicit K-dependence through the support size and new-class probability mass. The manuscript must supply the explicit calculation demonstrating that this dependence cancels for arbitrary K; without it the radius grows with K exactly as the Hoeffding bounds the authors criticize, undermining both the regularizer and the downstream inheritance guarantees.

Authors: We thank the referee for this observation. In deriving the McDiarmid radius for Gini impurity in §3.1, the per-sample bounded differences c_i are computed from the current class probabilities p_j. However, the sum of squared differences that enters the concentration radius simplifies such that all explicit K-dependent factors cancel, leaving a bound whose leading term depends only on the sample size n and a universal constant (arising because Gini lies in [0,1] and its one-sample sensitivity is at most 1). We will insert the complete algebraic expansion of this cancellation, including the limiting case of vanishing new-class probability mass, directly into the revised §3.1 so that readers can verify K-independence for arbitrary K. revision: yes
Referee: [§4.2] §4.2 (Bayesian Inheritance Protocol): The truncated-Gaussian moment projection claims variance reduction that is strongest precisely when splitting is most conservative. In the class-incremental regime, new classes enter after splits have occurred; the Gaussian assumption may then be violated, introducing bias that accumulates with K. A formal bias bound or empirical verification on streams with deliberately increasing K is needed to confirm the protocol does not degrade the claimed robustness.

Authors: We agree that the truncated-Gaussian moment projection can be violated once new classes appear after a split. The protocol exactly matches the first two moments of the parent distribution to the children, which guarantees variance reduction under the projection regardless of higher-order moments, but does not preclude bias from distribution mismatch. In the revision we will add a dedicated experimental subsection that incrementally introduces new classes at controlled points in synthetic and real streams, reporting tree depth, split quality, and accuracy trajectories with and without the inheritance step. We will also expand the discussion to note the limitation of the Gaussian assumption. A fully general, distribution-free bias bound is not provided in the current manuscript and would require additional assumptions; the new empirical results will therefore serve as the primary supporting evidence. revision: partial

Circularity Check

0 steps flagged

No circularity: McDiarmid bound and components are externally grounded

full rationale

The paper's core derivation applies the standard McDiarmid inequality (an external result) to obtain a Gini confidence radius claimed to be K-independent, without reducing the radius to a fitted parameter or self-derived quantity by construction. The Bayesian inheritance protocol and KLL sketches are presented as integrations of known techniques rather than tautological renamings or self-citations. No load-bearing self-citation chains, uniqueness theorems from the authors, or ansatzes smuggled via prior work appear in the provided abstract or claimed components. The derivation chain remains self-contained against external benchmarks like McDiarmid's inequality and standard quantile sketches.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

Abstract-only view limits visibility; the method rests on the McDiarmid inequality applied to Gini, an implicit assumption that data streams permit truncated-Gaussian moment matching, and the existence of KLL quantile sketches as a black-box primitive. No explicit free parameters or invented entities are named.

axioms (2)

domain assumption McDiarmid's inequality can be instantiated on the Gini impurity estimator to produce a radius independent of class count K
Invoked to replace Hoeffding-style bounds whose radius grows with log K
domain assumption Parent-to-child statistic transfer can be accurately modelled by moments of a truncated Gaussian distribution
Used in the Bayesian inheritance protocol

pith-pipeline@v0.9.0 · 5591 in / 1645 out tokens · 53898 ms · 2026-05-13T01:15:42.621195+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear
Theorem 5.1: ci ≤ 4/n for Gini split gain F(S), tight for arbitrary K; Corollary 4.1: ε = sqrt(32 ln(2dm/δ)/n) K-independent radius

Reference graph

Works this paper leans on

73 extracted references · 73 canonical work pages · 1 internal anchor

[1]

Memory aware synapses: Learning what (not) to forget

Rahaf Aljundi, Francesca Babiloni, Mohamed Elhoseiny, Marcus Rohrbach, and Tinne Tuyte- laars. Memory aware synapses: Learning what (not) to forget. InProceedings of the European Conference on Computer Vision (ECCV), September 2018

work page 2018
[2]

Online continual learning with maximal interfered retrieval

Rahaf Aljundi, Eugene Belilovsky, Tinne Tuytelaars, Laurent Charlin, Massimo Caccia, Min Lin, and Lucas Page-Caccia. Online continual learning with maximal interfered retrieval. In H. Wallach, H. Larochelle, A. Beygelzimer, F. d'Alché-Buc, E. Fox, and R. Garnett, editors, Advances in Neural Information Processing Systems, volume 32. Curran Associates, Inc., 2019

work page 2019
[3]

Expert gate: Lifelong learning with a network of experts

Rahaf Aljundi, Punarjay Chakravarty, and Tinne Tuytelaars. Expert gate: Lifelong learning with a network of experts. In2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 7120–7129, 2017

work page 2017
[4]

Gradient based sample selection for online continual learning

Rahaf Aljundi, Min Lin, Baptiste Goujaud, and Yoshua Bengio. Gradient based sample selection for online continual learning. In H. Wallach, H. Larochelle, A. Beygelzimer, F. d'Alché-Buc, E. Fox, and R. Garnett, editors,Advances in Neural Information Processing Systems, volume 32. Curran Associates, Inc., 2019

work page 2019
[5]

Il2m: Class incremental learning with dual memory

Eden Belouadah and Adrian Popescu. Il2m: Class incremental learning with dual memory. In 2019 IEEE/CVF International Conference on Computer Vision (ICCV), pages 583–592, 2019

work page 2019
[6]

A streaming parallel decision tree algorithm.J

Yael Ben-Haim and Elad Tom-Tov. A streaming parallel decision tree algorithm.J. Mach. Learn. Res., 11:849–872, March 2010

work page 2010
[7]

Learning from time-changing data with adaptive windowing

Albert Bifet and Ricard Gavaldà. Learning from time-changing data with adaptive windowing. volume 7, 04 2007

work page 2007
[8]

New ensemble methods for evolving data streams

Albert Bifet, Geoff Holmes, Bernhard Pfahringer, Richard Kirkby, and Ricard Gavaldà. New ensemble methods for evolving data streams. InProceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD ’09, page 139–148, New York, NY , USA, 2009. Association for Computing Machinery

work page 2009
[9]

Dark experience for general continual learning: a strong, simple baseline

Pietro Buzzega, Matteo Boschini, Angelo Porrello, Davide Abati, and Simone Calderara. Dark experience for general continual learning: a strong, simple baseline. InProceedings of the 34th International Conference on Neural Information Processing Systems, NIPS ’20, Red Hook, NY , USA, 2020. Curran Associates Inc

work page 2020
[10]

New insights on reducing abrupt representation change in online continual learning

Lucas Caccia, Rahaf Aljundi, Nader Asadi, Tinne Tuytelaars, Joelle Pineau, and Eugene Belilovsky. New insights on reducing abrupt representation change in online continual learning. InInternational Conference on Learning Representations (ICLR), 2022

work page 2022
[11]

Online learning of decision trees with Thompson sampling

Ayman Chaouki, Jesse Read, and Albert Bifet. Online learning of decision trees with Thompson sampling. In Sanjoy Dasgupta, Stephan Mandt, and Yingzhen Li, editors,Proceedings of The 27th International Conference on Artificial Intelligence and Statistics, volume 238 of Proceedings of Machine Learning Research, pages 2944–2952. PMLR, 02–04 May 2024

work page 2024
[12]

Dokania, Thalaiyasingam Ajanthan, and Philip H

Arslan Chaudhry, Puneet K. Dokania, Thalaiyasingam Ajanthan, and Philip H. S. Torr. Rie- mannian walk for incremental learning: Understanding forgetting and intransigence. In Vittorio Ferrari, Martial Hebert, Cristian Sminchisescu, and Yair Weiss, editors,Computer Vision – ECCV 2018, pages 556–572, Cham, 2018. Springer International Publishing

work page 2018
[13]

Efficient lifelong learning with A-GEM.CoRR, abs/1812.00420, 2018

Arslan Chaudhry, Marc’Aurelio Ranzato, Marcus Rohrbach, and Mohamed Elhoseiny. Efficient lifelong learning with A-GEM.CoRR, abs/1812.00420, 2018

work page arXiv 2018
[14]

Arslan Chaudhry, Marcus Rohrbach, Mohamed Elhoseiny, Thalaiyasingam Ajanthan, Puneet Ku- mar Dokania, Philip H. S. Torr, and Marc’Aurelio Ranzato. Continual learning with tiny episodic memories.CoRR, abs/1902.10486, 2019

work page arXiv 1902
[15]

A continual learning survey: Defying forgetting in classification tasks.IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(7):3366–3385, 2022

Matthias De Lange, Rahaf Aljundi, Marc Masana, Sarah Parisot, Xu Jia, Aleš Leonardis, Gregory Slabaugh, and Tinne Tuytelaars. A continual learning survey: Defying forgetting in classification tasks.IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(7):3366–3385, 2022. 11

work page 2022
[16]

Splitting with confidence in decision trees with application to stream mining

Rocco De Rosa and Nicolò Cesa-Bianchi. Splitting with confidence in decision trees with application to stream mining. In2015 International Joint Conference on Neural Networks (IJCNN), pages 1–8, 2015

work page 2015
[17]

Mining high-speed data streams

Pedro Domingos and Geoff Hulten. Mining high-speed data streams. InProceedings of the Sixth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD ’00, page 71–80, New York, NY , USA, 2000. Association for Computing Machinery

work page 2000
[18]

Robert M. French. Catastrophic forgetting in connectionist networks.Trends in Cognitive Sciences, 3(4):128–135, 1999

work page 1999
[19]

João Gama, Pedro Medas, Gladys Castillo, and Pedro Rodrigues.Learning with Drift Detection, volume 8, pages 286–295. 09 2004

work page 2004
[20]

Accurate decision trees for mining high-speed data streams

João Gama, Ricardo Rocha, and Pedro Medas. Accurate decision trees for mining high-speed data streams. pages 523–528, 08 2003

work page 2003
[21]

Díaz-Redondo

Pablo García-Santaclara, Bruno Fernández-Castro, and Rebeca P. Díaz-Redondo. Overcom- ing catastrophic forgetting in tabular data classification: A pseudorehearsal-based approach. Engineering Applications of Artificial Intelligence, 156:110908, 2025

work page 2025
[22]

Torr, and Bernard Ghanem

Yasir Ghunaim, Adel Bibi, Kumail Alhamoud, Motasem Alfarra, Hasan Abed Al Kader Ham- moud, Ameya Prabhu, Philip H.S. Torr, and Bernard Ghanem. Real-time evaluation in online continual learning: A new hope. In2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 11888–11897, 2023

work page 2023
[23]

Continual contrastive learning on tabular data with out of distribution

Achmad Ginanjar, Xue Li, Priyanka Singh, and Wen Hua. Continual contrastive learning on tabular data with out of distribution. InESANN 2025 proceedings, ESANN 2025, page 93–98. Ciaco - i6doc.com, 2025

work page 2025
[24]

Adaptive random forests for evolving data stream classification.Machine Learning, 106:1–27, 10 2017

Heitor Murilo Gomes, Albert Bifet, Jesse Read, Jean Paul Barddal, Fabrício Enembreck, Bernhard Pfahringer, Geoff Holmes, and Talel Abdessalem. Adaptive random forests for evolving data stream classification.Machine Learning, 106:1–27, 10 2017

work page 2017
[25]

An empirical investigation of catastrophic forgetting in gradient-based neural networks.arXiv preprint arXiv:1312.6211,

Ian Goodfellow, Mehdi Mirza, Xia Da, and Aaron Courville. An empirical investigation of catastrophic forgetting in gradient-based neural networks.arXiv preprint arXiv:1312.6211, 12 2013

work page arXiv 2013
[26]

Fecam: Exploiting the heterogeneity of class distributions in exemplar-free continual learning

Dipam Goswami, Yuyang Liu, Bartłomiej Twardowski, and Joost van de Weijer. Fecam: Exploiting the heterogeneity of class distributions in exemplar-free continual learning. In A. Oh, T. Naumann, A. Globerson, K. Saenko, M. Hardt, and S. Levine, editors,Advances in Neural Information Processing Systems, volume 36, pages 6582–6595. Curran Associates, Inc., 2023

work page 2023
[27]

Hayes, Kushal Kafle, Robik Shrestha, Manoj Acharya, and Christopher Kanan

Tyler L. Hayes, Kushal Kafle, Robik Shrestha, Manoj Acharya, and Christopher Kanan. Remind your neural network to prevent catastrophic forgetting. In Andrea Vedaldi, Horst Bischof, Thomas Brox, and Jan-Michael Frahm, editors,Computer Vision – ECCV 2020, pages 466–483, Cham, 2020. Springer International Publishing

work page 2020
[28]

Hayes and Christopher Kanan

Tyler L. Hayes and Christopher Kanan. Lifelong machine learning with deep streaming linear discriminant analysis. In2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), pages 887–896, 2020

work page 2020
[29]

Continual learning for unsupervised anomaly detection in continuous auditing of financial accounting data, 2022

Hamed Hemati, Marco Schreyer, and Damian Borth. Continual learning for unsupervised anomaly detection in continuous auditing of financial accounting data, 2022

work page 2022
[30]

Time-uniform, nonparametric, nonasymptotic confidence sequences.The Annals of Statistics, 49, 04 2021

Steven Howard, Aaditya Ramdas, Jon McAuliffe, and Jagmohan Sekhon. Time-uniform, nonparametric, nonasymptotic confidence sequences.The Annals of Statistics, 49, 04 2021

work page 2021
[31]

Re-evaluating continual learning scenarios: A categorization and case for strong baselines, 2019

Yen-Chang Hsu, Yen-Cheng Liu, Anita Ramasamy, and Zsolt Kira. Re-evaluating continual learning scenarios: A categorization and case for strong baselines, 2019

work page 2019
[32]

Mining time-changing data streams

Geoff Hulten, Laurie Spencer, and Pedro Domingos. Mining time-changing data streams. In Proceedings of the Seventh ACM SIGKDD International Conference on Knowledge Discov- ery and Data Mining, KDD ’01, page 97–106, New York, NY , USA, 2001. Association for Computing Machinery. 12

work page 2001
[33]

Selective experience replay for lifelong learning

David Isele and Akansel Cosgun. Selective experience replay for lifelong learning. AAAI’18/IAAI’18/EAAI’18. AAAI Press, 2018

work page 2018
[34]

Estimating continuous distributions in bayesian classifiers

George John and Pat Langley. Estimating continuous distributions in bayesian classifiers. Proceedings of the 11th Conference on Uncertainty in Artificial Intelligence, 1, 02 2013

work page 2013
[35]

Optimal quantile approximation in streams

Zohar Karnin, Kevin Lang, and Edo Liberty. Optimal quantile approximation in streams. In 2016 IEEE 57th Annual Symposium on Foundations of Computer Science (FOCS), pages 71–78, 2016

work page 2016
[36]

Rusu, Kieran Milan, John Quan, Tiago Ramalho, Agnieszka Grabska-Barwinska, Demis Hassabis, Claudia Clopath, Dharshan Kumaran, and Raia Hadsell

James Kirkpatrick, Razvan Pascanu, Neil Rabinowitz, Joel Veness, Guillaume Desjardins, Andrei A. Rusu, Kieran Milan, John Quan, Tiago Ramalho, Agnieszka Grabska-Barwinska, Demis Hassabis, Claudia Clopath, Dharshan Kumaran, and Raia Hadsell. Overcoming catas- trophic forgetting in neural networks.Proceedings of the National Academy of Sciences, 114(13):352...

work page 2017
[37]

Class-incremental experience replay for continual learning under concept drift

Łukasz Korycki and Bartosz Krawczyk. Class-incremental experience replay for continual learning under concept drift. In2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), pages 3644–3653, 2021

work page 2021
[38]

Continual learning for robotics: Definition, framework, learning strategies, opportunities and challenges.Inf

Timothée Lesort, Vincenzo Lomonaco, Andrei Stoian, Davide Maltoni, David Filliat, and Natalia Díaz-Rodríguez. Continual learning for robotics: Definition, framework, learning strategies, opportunities and challenges.Inf. Fusion, 58(C):52–68, June 2020

work page 2020
[39]

Learning without forgetting.IEEE Transactions on Pattern Analysis and Machine Intelligence, 40(12):2935–2947, 2018

Zhizhong Li and Derek Hoiem. Learning without forgetting.IEEE Transactions on Pattern Analysis and Machine Intelligence, 40(12):2935–2947, 2018

work page 2018
[40]

Gradient episodic memory for continual learning

David Lopez-Paz and Marc’Aurelio Ranzato. Gradient episodic memory for continual learning. InProceedings of the 31st International Conference on Neural Information Processing Systems, NIPS’17, page 6470–6479, Red Hook, NY , USA, 2017. Curran Associates Inc

work page 2017
[41]

Online continual learning in image classification: An empirical survey.Neurocomputing, 469:28–51, 2022

Zheda Mai, Ruiwen Li, Jihwan Jeong, David Quispe, Hyunwoo Kim, and Scott Sanner. Online continual learning in image classification: An empirical survey.Neurocomputing, 469:28–51, 2022

work page 2022
[42]

Piggyback: Adapting a single network to multiple tasks by learning to mask weights

Arun Mallya, Dillon Davis, and Svetlana Lazebnik. Piggyback: Adapting a single network to multiple tasks by learning to mask weights. InComputer Vision – ECCV 2018: 15th European Conference, Munich, Germany, September 8-14, 2018, Proceedings, Part IV, page 72–88, Berlin, Heidelberg, 2018. Springer-Verlag

work page 2018
[43]

PackNet: Adding Multiple Tasks to a Single Network by Iterative Pruning

Arun Mallya and Svetlana Lazebnik. PackNet: Adding Multiple Tasks to a Single Network by Iterative Pruning . In2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 7765–7773, Los Alamitos, CA, USA, June 2018. IEEE Computer Society

work page 2018
[44]

Webb, and Mahsa Salehi

Chaitanya Manapragada, Geoffrey I. Webb, and Mahsa Salehi. Extremely fast decision tree. InProceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, KDD ’18, page 1953–1962, New York, NY , USA, 2018. Association for Computing Machinery

work page 1953
[45]

Michael McCloskey and Neal J. Cohen. Catastrophic interference in connectionist networks: The sequential learning problem. volume 24 ofPsychology of Learning and Motivation, pages 109–165. Academic Press, 1989

work page 1989
[46]

Distance-based image classification: Generalizing to new classes at near-zero cost.IEEE Transactions on Pattern Analysis and Machine Intelligence, 35:2624–37, 11 2013

Thomas Mensink, Jakob Verbeek, and Gabriela Csurka. Distance-based image classification: Generalizing to new classes at near-zero cost.IEEE Transactions on Pattern Analysis and Machine Intelligence, 35:2624–37, 11 2013

work page 2013
[47]

Boosted dyadic kernel discriminants

Baback Moghaddam and Gregory Shakhnarovich. Boosted dyadic kernel discriminants. In S. Becker, S. Thrun, and K. Obermayer, editors,Advances in Neural Information Processing Systems, volume 15. MIT Press, 2002

work page 2002
[48]

Parisi, Ronald Kemker, Jose L

German I. Parisi, Ronald Kemker, Jose L. Part, Christopher Kanan, and Stefan Wermter. Continual lifelong learning with neural networks: A review.Neural Networks, 113:54–71, 2019. 13

work page 2019
[49]

Latent replay for real-time continual learning

Lorenzo Pellegrini, Gabriele Graffieti, Vincenzo Lomonaco, and Davide Maltoni. Latent replay for real-time continual learning. In2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), page 10203–10209. IEEE Press, 2020

work page 2020
[50]

Fetril: Feature translation for exemplar-free class-incremental learning

Grégoire Petit, Adrian Popescu, Hugo Schindler, David Picard, and Bertrand Delezoide. Fetril: Feature translation for exemplar-free class-incremental learning. InProceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), pages 3911– 3920, January 2023

work page 2023
[51]

Ameya Prabhu, Philip H. S. Torr, and Puneet K. Dokania. Gdumb: A simple approach that questions our progress in continual learning. InComputer Vision – ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part II, page 524–540, Berlin, Heidelberg, 2020. Springer-Verlag

work page 2020
[52]

Connectionist models of recognition memory: Constraints imposed by learning and forgetting functions.Psychological Review, 97:285–308, 04 1990

Roger Ratcliff. Connectionist models of recognition memory: Constraints imposed by learning and forgetting functions.Psychological Review, 97:285–308, 04 1990

work page 1990
[53]

Sylvestre-Alvise Rebuffi, Alexander Kolesnikov, Georg Sperl, and Christoph H. Lampert. icarl: Incremental classifier and representation learning. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), July 2017

work page 2017
[54]

Lillicrap, and Greg Wayne.Experi- ence replay for continual learning

David Rolnick, Arun Ahuja, Jonathan Schwarz, Timothy P. Lillicrap, and Greg Wayne.Experi- ence replay for continual learning. Curran Associates Inc., Red Hook, NY , USA, 2019

work page 2019
[55]

Progressive Neural Networks

Andrei Rusu, Neil Rabinowitz, Guillaume Desjardins, Hubert Soyer, James Kirkpatrick, Koray Kavukcuoglu, Razvan Pascanu, and Raia Hadsell. Progressive neural networks.arXiv preprint arXiv:1606.04671, 06 2016

work page internal anchor Pith review Pith/arXiv arXiv 2016
[56]

Decision trees for mining data streams based on the mcdiarmid’s bound.IEEE Transactions on Knowledge and Data Engineering, 25(6):1272–1279, 2013

Leszek Rutkowski, Lena Pietruczuk, Piotr Duda, and Maciej Jaworski. Decision trees for mining data streams based on the mcdiarmid’s bound.IEEE Transactions on Knowledge and Data Engineering, 25(6):1272–1279, 2013

work page 2013
[57]

Progress & compress: A scalable framework for continual learning

Jonathan Schwarz, Wojciech Czarnecki, Jelena Luketina, Agnieszka Grabska-Barwinska, Yee Whye Teh, Razvan Pascanu, and Raia Hadsell. Progress & compress: A scalable framework for continual learning. In Jennifer Dy and Andreas Krause, editors,Proceedings of the 35th International Conference on Machine Learning, volume 80 ofProceedings of Machine Learning Re...

work page 2018
[58]

Prototype reminiscence and augmented asymmetric knowledge aggregation for non-exemplar class-incremental learning

Wuxuan Shi and Mang Ye. Prototype reminiscence and augmented asymmetric knowledge aggregation for non-exemplar class-incremental learning. In2023 IEEE/CVF International Conference on Computer Vision (ICCV), pages 1772–1781, 2023

work page 2023
[59]

Continual learning with deep gener- ative replay

Hanul Shin, Jung Kwon Lee, Jaehong Kim, and Jiwon Kim. Continual learning with deep gener- ative replay. In I. Guyon, U. V on Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett, editors,Advances in Neural Information Processing Systems, volume 30. Curran Associates, Inc., 2017

work page 2017
[60]

CODA-Prompt: COntinual Decomposed Attention-Based Prompting for Rehearsal-Free Continual Learning

James Seale Smith, Leonid Karlinsky, Vyshnavi Gutta, Paola Cascante-Bonilla, Donghyun Kim, Assaf Arbelle, Rameswar Panda, Rogerio Feris, and Zsolt Kira. CODA-Prompt: COntinual Decomposed Attention-Based Prompting for Rehearsal-Free Continual Learning . In2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 11909– 11919, Los Al...

work page 2023
[61]

Three types of incremental learning

Gido van de Ven, Tinne Tuytelaars, and Andreas Tolias. Three types of incremental learning. Nature Machine Intelligence, 4:1–13, 12 2022

work page 2022
[62]

van de Ven and Andreas S

Gido M. van de Ven and Andreas S. Tolias. Three scenarios for continual learning.CoRR, abs/1904.07734, 2019

work page arXiv 1904
[63]

An attention-based feature memory design for energy-efficient continual learning, 2025

Yuandou Wang, Filip Gunnarsson, and Rihan Hai. An attention-based feature memory design for energy-efficient continual learning, 2025. 14

work page 2025
[64]

Dualprompt: Complementary prompting for rehearsal-free continual learning

Zifeng Wang, Zizhao Zhang, Sayna Ebrahimi, Ruoxi Sun, Han Zhang, Chen-Yu Lee, Xiaoqi Ren, Guolong Su, Vincent Perot, Jennifer Dy, and Tomas Pfister. Dualprompt: Complementary prompting for rehearsal-free continual learning. In Shai Avidan, Gabriel Brostow, Moustapha Cissé, Giovanni Maria Farinella, and Tal Hassner, editors,Computer Vision – ECCV 2022, pag...

work page 2022
[65]

Learning to prompt for continual learning

Zifeng Wang, Zizhao Zhang, Chen-Yu Lee, Han Zhang, Ruoxi Sun, Xiaoqi Ren, Guolong Su, Vincent Perot, Jennifer Dy, and Tomas Pfister. Learning to prompt for continual learning. In2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 139–149, 2022

work page 2022
[66]

Lifelong learning with dynamically expandable networks

Jaehong Yoon, Eunho Yang, Jeongtae Lee, and Sung Ju Hwang. Lifelong learning with dynamically expandable networks. InInternational Conference on Learning Representations (ICLR), 2018

work page 2018
[67]

Prediction error-based classification for class-incremental learning

Michał Zaj ˛ ac, Tinne Tuytelaars, and Gido van de Ven. Prediction error-based classification for class-incremental learning. In B. Kim, Y . Yue, S. Chaudhuri, K. Fragkiadaki, M. Khan, and Y . Sun, editors,International Conference on Learning Representations, volume 2024, pages 8239–8267, 2024

work page 2024
[68]

Continual learning through synaptic intelligence

Friedemann Zenke, Ben Poole, and Surya Ganguli. Continual learning through synaptic intelligence. In Doina Precup and Yee Whye Teh, editors,Proceedings of the 34th International Conference on Machine Learning, volume 70 ofProceedings of Machine Learning Research, pages 3987–3995. PMLR, 06–11 Aug 2017

work page 2017
[69]

Prototype augmentation and self-supervision for incremental learning

Fei Zhu, Xu-Yao Zhang, Chuang Wang, Fei Yin, and Cheng-Lin Liu. Prototype augmentation and self-supervision for incremental learning. In2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 5867–5876, 2021

work page 2021
[70]

Self-sustaining representation expansion for non-exemplar class-incremental learning

Kai Zhu, Wei Zhai, Yang Cao, Jiebo Luo, and Zheng-Jun Zha. Self-sustaining representation expansion for non-exemplar class-incremental learning. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 9296–9305, June 2022

work page 2022
[71]

Gacl: exemplar-free generalized analytic continual learning

Huiping Zhuang, Yizhu Chen, Di Fang, Run He, Kai Tong, Hongxin Wei, Ziqian Zeng, and Cen Chen. Gacl: exemplar-free generalized analytic continual learning. InProceedings of the 38th International Conference on Neural Information Processing Systems, NIPS ’24, Red Hook, NY , USA, 2024. Curran Associates Inc

work page 2024
[72]

1− ζϕ(ζ) ˜Φ − ϕ(ζ) ˜Φ 2# {upper-truncated; Appendix I.1} 9:else 10:(σ s,c j∗ )2 ←σ c2 j∗

Huiping Zhuang, Zhenyu Weng, Hongxin Wei, Renchuzi Xie, Kar-Ann Toh, and Zhiping Lin. Acil: Analytic class-incremental learning with absolute memorization and privacy protection. In S. Koyejo, S. Mohamed, A. Agarwal, D. Belgrave, K. Cho, and A. Oh, editors,Advances in Neural Information Processing Systems, volume 35, pages 11602–11614. Curran Associates, ...

work page arXiv 2022
[73]

IRB approval is not applicable

Institutional review board (IRB) approvals or equivalent for research with human subjects Question: Does the paper describe potential risks incurred by study participants, whether such risks were disclosed to the subjects, and whether Institutional Review Board (IRB) approvals (or an equivalent approval/review based on the requirements of your country or ...

work page