Pith. sign in

REVIEW 4 major objections 5 minor 42 references

Improving Model Classification by Optimizing the Training Dataset

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper claims that tuning how a coreset is sampled—adding deterministic inclusion, class-wise budgets, and active refinement—produces small training sets whose models beat both vanilla coresets and, in most cases, full-dataset training.

desk verdict An honest empirical survey of tuning coreset sampling for classification metrics, but the no-error-bars grid-search selection and untuned full baselines blow a hole in the 'tuned beats full' claim. read the letter →

arxiv 2507.16729 v1 pith:7Z524MWS submitted 2025-07-22 cs.LG

Morad Tukan , Loay Mualem , Eitan Netzer , Liran Sigalat This is my paper · ORCID
classification cs.LG
keywords coresetsdata-centricAItrainingsetselectionimportancesamplingclassificationmetricsF1scoreactiveclassimbalance
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

Coresets are weighted subsets of a training set chosen to reproduce the full-data training loss, and the paper's starting point is that reproducing the loss is not the same as training a good classifier. The authors argue that the coreset recipe itself can be tuned for the classifier you actually want, and they introduce three adjustable mechanisms: deterministically including the highest-probability points to restore diversity, allocating the sample budget across classes by hand instead of by class frequency, and refining the subset with active sampling that queries the points the current model is most unsure about. Across logistic regression, SVM, and XGBoost on eight datasets, the tuned coresets outperform vanilla coresets on metrics such as $F1$ and balanced accuracy, and in most experiments they also outperform models trained on the full data. If the results hold, data curation becomes a tunable part of model design: a smaller training set can be better than the full one, not just cheaper.

What carries the argument

The central object is the *tuned coreset*, built on a sensitivity-based $\varepsilon$-coreset (a weighted subset drawn by importance sampling with probabilities derived from sensitivity scores) and modified by three mechanisms. Deterministic sampling takes the highest-probability points into the coreset exactly once and reweights the rest, following the diversity-preserving construction and Theorem 31 of [FSS20]. Class-wise allocation splits the sample budget across classes by a tunable ratio rather than the data's natural class frequencies. Active sampling (Algorithm 1) repeatedly trains the model on the current coreset, asks an active learner for the $m$ points the model is most uncertain about from the remaining training data, and adds them until a patience counter sees no validation-metric improvement; the algorithm returns whichever of the original or refined coreset scores better. Proposition 1—that a logistic-regression coreset is practically sufficient for decision-tree classification, justified by gradient boosting implementing logistic regression additively—is the bridge that lets the unified coreset of [TMF20] drive the XGBoost experiments.

What would settle it

Train XGBoost on the same datasets with (a) the unified logistic-regression coreset of [TMF20], (b) an equal-size coreset built from a loss native to trees, such as the MSE-based decision-tree coreset of [JSSNF21], and (c) uniform sampling; if (a) does not match or beat (b) and (c) on validation and test $F1$, then Proposition 1 is false and the Section 4.4 active-sampling gains reduce to generic subset selection rather than coreset guarantees.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that sensitivity-based coreset construction has hidden tunable knobs—deterministic sampling, class-wise sample allocation, and deterministic weight-handling strategies (keep, inv, prop)—and that turning them with a validation-set grid search materially changes downstream classification quality. Tuned coresets are reported to beat their vanilla versions by roughly 2–3 percentage points or more in $F1$ score across logistic regression and SVM on the A9A and CodRNA benchmarks, and to match or exceed full-data training in most configurations. For XGBoost, active sampling extends the same gains to strongly imbalanced fraud-detection datasets (IEEE, Credit Card, E-commerce), where the refined coreset beats the full-data model on balanced accuracy and $F1$. The paper frames this as evidence that a coreset can be an opportunity to improve generalization and mitigate class bias, not merely an approximation device.

Load-bearing premise

The XGBoost results depend on Proposition 1: that a coreset built for logistic regression is practically sufficient for decision-tree classification, which the paper supports by analogy to gradient boosting rather than by proof; if that bridge fails, the tree-boosting results lose their coreset grounding.

Editorial extensions

If this is right

  • Coreset size can be spent deliberately: shifting part of the sample budget toward minority classes improves balanced accuracy and $F1$ without changing the total subset size, and the paper notes this does not harm the $\varepsilon$-coreset approximation guarantee.
  • Where vanilla sensitivity sampling is most degenerate—when a few points dominate the sampling probabilities—deterministic inclusion restores diversity, and the weight-handling strategies (keep, inv, prop) give a tunable trade-off between the deterministic and probabilistic parts of the coreset.
  • Active sampling converts a coreset from a static compression into an iterative refinement loop that, by construction, returns the better of the original or refined coreset with respect to the chosen metric on the validation set.
  • For tree ensembles, the results imply that a coreset designed for logistic loss can serve as a practical training subset for XGBoost, beating uniform sampling and, on imbalanced fraud data, beating full-data training on balanced accuracy and $F1$.

Reading between the lines

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

  • Implicit in the method is that coreset construction becomes a hyperparameter-optimization problem; the authors flag meta-learning or reinforcement learning for picking the sampling parameters as future work, which would make the tuning loop itself automatic.
  • If the central claim generalizes, the training value of a point is not uniform: minority-class points, points near the decision boundary, and high-sensitivity outliers carry more value per sample than the bulk, so any curation method that ignores class balance and diversity forfeits classification performance.
  • A testable extension the paper does not run is multiclass or regression analogues; if the same three knobs do not produce similar relative gains there, the effect may be specific to binary classification with $F1$-style metrics.
  • Because uniform random sampling is sometimes competitive with vanilla sensitivity coresets in the reported tables, a separate ablation that applies the same three knobs to a random (non-sensitivity) coreset would separate how much of the gain comes from the knobs and how much from the sensitivity theory.
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 / 5 minor

Summary. The paper proposes a practical framework for tuning coreset construction for classification tasks. It introduces three mechanisms on top of sensitivity-based coresets: deterministic inclusion of high-probability points with reweighting, class-wise allocation of the sample budget, and an active-sampling refinement loop (Algorithm 1). The authors report experiments on eight binary datasets using logistic regression, SVM, and XGBoost, and claim that tuned coresets significantly outperform vanilla coresets and often outperform training on the full dataset with respect to balanced accuracy, F1, and ROC AUC.

Significance. If the central claim holds, the paper would make a useful contribution to data-centric machine learning by showing that coreset generation can be optimized for downstream classification metrics rather than for loss approximation alone. The breadth of datasets and classifiers, and the ablation study that isolates each tunable parameter, are strengths. However, the main claim is not yet established by the evidence as presented: the evaluation protocol selects the best validation configuration over a large grid, the full-data baselines are not given an equivalent hyperparameter search, no error bars or repeated-seed results are reported, and the decision-tree experiments rely on an unproven proposition. With the current evidence, the paper is best read as a preliminary empirical report rather than a definitive demonstration.

major comments (4)
  1. [§4.1–4.3, Tables 1–3, Appendix A] The 'tuned' configurations are selected by taking the argmax of validation F1 over a grid of roughly 5 coreset sizes × 7 class allocations × 6 deterministic ratios × 3 weight schemes (Tables 6–7), but the paper reports no error bars, no repeated sampling seeds, and no distribution of test metrics over the grid. Moreover, the full-data baselines in Sections 4.1–4.3 use a fixed regularization parameter (Section 4.1: 'a regularization parameter equal to 1') and XGBoost library defaults, so the 'tuned beats full' comparison conflates data tuning with model selection. Because the maximum of many noisy estimates is biased upward even when the tuning signal is pure noise, the headline claim in the abstract is not supported unless the authors report the full grid distribution, repeat each configuration over seeds, and compare against a full-data baseline that receives an equivalent hyperparameter search. The results in Table 5, where both sides are tuned, show the advantage shrinking or reversing on several metrics (e.g., IEEE F1 74.34 vs 74.56; IEEE ROC AUC 95.81 vs 96.59), which is consistent with this concern.
  2. [§4.3, Proposition 1] Proposition 1 states that a coreset for logistic regression is sufficient for decision tree classification 'from a practical point of view,' but no proof is given, and the cited fact that gradient boosting implements logistic regression additively does not imply that an ε-coreset for the logistic loss preserves the XGBoost training objective or the resulting classification metrics. Since the entire DTC section depends on this transfer, either replace the proposition with a precise formal statement and proof, or reclassify it as an empirical conjecture and weaken the claims accordingly.
  3. [Table 4, Hepmass rows] The validation ROC AUC values of 3.77%, 3.75%, 3.64%, and 2.72% are not valid ROC AUC values for a binary classifier on a balanced dataset; an ROC AUC below 50% (and especially near 3%) is implausible unless the quantity being reported is something else. The test ROC AUC values in the same table are approximately 70%, so the table is internally inconsistent. This must be corrected, and the reported metric must be defined precisely, because Table 4 is one of the main pieces of evidence for the DTC claim.
  4. [§3, Algorithm 1] Algorithm 1's output is described as 'an ε-coreset ... that practically outperforms' the input coreset, but no argument shows that adding actively sampled points with weight 1 preserves the ε-coreset property; the 'better or same' property is only a consequence of the final validation comparison in Lines 18–22, not of coreset theory. The active sampling query is also unspecified: the manuscript does not name the active learning algorithm, the query size m, the patience ρ, or the final coreset sizes used in Table 5. Please specify these details and either prove or remove the ε-coreset guarantee from the output description.
minor comments (5)
  1. [§2.1] The notation paragraph is incomplete: 'let Y denote the set of functions mapping instances from X to h' ends abruptly and never defines h or Y; please fix.
  2. [Appendix C, Figures 5–10] Several figure captions list '(b)' twice (e.g., 'where (a) refers to using a coreset ratio of 0.5%, (b) 5.4%, (b) 10.3%, and (d) 20%'); the second '(b)' should be '(c)'.
  3. [§4.1, Table 1] The dataset names are used inconsistently: 'A9A' and 'cod-rna' appear alongside the names 'Adult' and 'CodRNA' used elsewhere; please unify the nomenclature.
  4. [§4.4, Table 5] No details are given for the active-sampling implementation in Table 5, including the specific query algorithm, the number of points m added per iteration, the patience ρ, the number of iterations until convergence, and the hyperparameter tuning protocol for the Full∗ and Active sampling∗ rows; these details are needed for reproducibility.
  5. [References, [Dat22]] The DataHeroes framework [Dat22] appears to be a core component of the implementation, but the reference is only a company name without a version, documentation URL, or repository; please provide sufficient information for reproducibility, and state any commercial interest.

Circularity Check

1 steps flagged · score 1.0 of 10

Only a minor definitional guarantee in Algorithm 1; test-set evidence is independent, so the central claim is not circular.

  1. self definitional [Section 3, 'Overview of Algorithm 1' and Algorithm 1, lines 18–23]
    "we output a coreset that can only be either better or of same performance quality to the given coreset concerning the classification metric. ... Finally, we return either the given coreset or its modified version based on which coreset leads to higher gain concerning the given classification metric ϕ (·) on the validation data."

    The guarantee is exactly the algorithm's return rule: Lines 18–23 train both the input coreset and the active-sampled version, compare ϕ on the validation data, and return whichever is higher. Hence 'better or same' is true by construction and carries no empirical content about active sampling. This is a self-definitional statement, but it is minor because the abstract's central claim is supported by the independent test-set metrics, not by this guarantee.

full rationale

The paper's central claim, that tuned coresets outperform vanilla coresets and often full-data training, is supported by test-set metrics that are external to the tuning procedure. The main weaknesses are experimental-design issues rather than circularity: the 'tuned' configuration is chosen as the validation-F1 argmax over a large grid, so validation improvements are partly a selection artifact and the full-data baseline is not comparably tuned; Proposition 1 is a stated practical heuristic validated by experiments rather than a proved theorem; and the theoretical backing for the DTC connection leans on the self-cited strong-coreset property of [TMF20]. None of these reduces a derived result to its own inputs: the test-set numbers are independent, and no equation in the paper is equivalent to a fitted value by construction. The only genuinely definitional element is Algorithm 1's 'better or same' guarantee, which follows directly from its return rule. That step is not load-bearing for the empirical claims, so the paper has no significant circularity.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The empirical claims depend on tuning parameters chosen by grid search; the theoretical guarantees are inherited from FSS20 and TMF20. The paper introduces no new entities, but it asserts two unproven premises: Proposition 1 and the active-sampling-preserves-coreset assumption.

free parameters (6)
  • class_size_allocation_ratio = selected by grid search per dataset (e.g., {0:65%,1:35%} for A9A DTC)
    Grid over class ratios in Tables 6-7; the best value on validation F1 is reported as tuned.
  • deterministic_ratio = selected per dataset from grids like 5-50% (SVM/LR) or 0.5-20% (DTC)
    Fraction of high-sensitivity points included deterministically; tuned on validation.
  • deterministic_weight_function = one of keep/inv/prop selected per experiment
    Weight-handling strategy chosen on validation.
  • active_sampling_query_size_m = not reported
    Algorithm 1 requires an input m (samples added per iteration), but experiments do not state its value.
  • patience_parameter_rho = not reported
    Algorithm 1 requires patience rho, but experiments do not state its value.
  • model_hyperparameters_after_tuning = not reported
    Section 4.4 says hyperparameter tuning was performed for DTC, but selected values are not listed.
assumptions (4)
  • standard math Sensitivity-based importance sampling plus deterministic inclusion following FSS20 Theorem 31 yields an epsilon-coreset.
    Section 2.2 relies on Theorem 31 of [FSS20] to justify adjusted sensitivities.
  • domain assumption Training, validation, and test splits are i.i.d. samples from the same distribution.
    Section 1 states each set is drawn i.i.d. from a distribution Sx.
  • ad hoc to paper Gradient boosting implements logistic regression additively, so a logistic regression coreset suffices for decision tree classification.
    Section 4.3, Proposition 1, asserted without proof and only empirically checked.
  • ad hoc to paper Adding actively sampled points to a coreset preserves the epsilon-coreset property.
    Algorithm 1 output is labeled an epsilon-coreset, but no proof is given for weight updates or guarantees.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Improving Model Classification by Optimizing the Training Dataset." pith.science (2026). https://pith.science/paper/7Z524MWS

@misc{pith2026250716729,
  author       = {Pith},
  title        = {Pith review of: Improving Model Classification by Optimizing the Training Dataset},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7Z524MWS}},
  note         = {Machine review of arXiv:2507.16729}
}
abstract

In the era of data-centric AI, the ability to curate high-quality training data is as crucial as model design. Coresets offer a principled approach to data reduction, enabling efficient learning on large datasets through importance sampling. However, conventional sensitivity-based coreset construction often falls short in optimizing for classification performance metrics, e.g., $F1$ score, focusing instead on loss approximation. In this work, we present a systematic framework for tuning the coreset generation process to enhance downstream classification quality. Our method introduces new tunable parameters--including deterministic sampling, class-wise allocation, and refinement via active sampling, beyond traditional sensitivity scores. Through extensive experiments on diverse datasets and classifiers, we demonstrate that tuned coresets can significantly outperform both vanilla coresets and full dataset training on key classification metrics, offering an effective path towards better and more efficient model training.

Figures

Figures reproduced from arXiv: 2507.16729 by the authors.

Figure 1
Figure 1. F1 scores across coreset sizes for multiple datasets (Logistic Regression). [PITH_FULL_IMAGE:figures/full_fig_p010_1.png] view at source ↗
Figure 2
Figure 2. F1 scores across coreset sizes for multiple datasets (SVM). [PITH_FULL_IMAGE:figures/full_fig_p012_2.png] view at source ↗
Figure 3
Figure 3. F1 scores across coreset sizes for multiple datasets (DTC). [PITH_FULL_IMAGE:figures/full_fig_p014_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Visualizations concerning Dataset (i). C Ablation study C.1 The effect of a single tunable coreset sampling parameter In what follows, we will inspect the effect of each tunable parameter in our data tun￾ing system on each of the following classification metrics: (i) A…
Figure 5
Figure 5. Figure 5: Ablation study on the effect of each tunable coreset sampling parameter on [PITH_FULL_IMAGE:figures/full_fig_p023_5.png]
Figure 6
Figure 6. Figure 6: Ablation study on the effect of each tunable coreset sampling parameter on [PITH_FULL_IMAGE:figures/full_fig_p024_6.png]
Figure 7
Figure 7. Figure 7: Ablation study on the effect of each tunable coreset sampling parameter on [PITH_FULL_IMAGE:figures/full_fig_p025_7.png]
Figure 8
Figure 8. Figure 8: Ablation study on the effect of each tunable coreset sampling parameter on [PITH_FULL_IMAGE:figures/full_fig_p026_8.png]
Figure 9
Figure 9. Figure 9: Ablation study on the effect of each tunable coreset sampling parameter on [PITH_FULL_IMAGE:figures/full_fig_p027_9.png]
Figure 10
Figure 10. Figure 10: Ablation study on the effect of each tunable coreset sampling parameter on [PITH_FULL_IMAGE:figures/full_fig_p028_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

42 extracted references · 39 canonical work pages

  1. [1]

    Physical Unclonable Functions

    Ahmad Aseeri, Mohammed Alkatheiri, and Yu Zhuang. Physical Unclonable Functions . UCI Machine Learning Repository, 2018. DOI : https://doi.org/10.24432/C5D03R

    work page doi:10.24432/c5d03r 2018

  2. [2]

    The power of uniform sampling for coresets

    Vladimir Braverman, Vincent Cohen-Addad, H-C Shaofeng Jiang, Robert Krauthgamer, Chris Schwiegelshohn, Mads Bech Toftrup, and Xuan Wu. The power of uniform sampling for coresets. In 2022 IEEE 63rd Annual Symposium on Foundations of Computer Science (FOCS) , pages 462--473. IEEE, 2022

    work page 2022

  3. [3]

    New frameworks for offline and streaming coreset constructions

    Vladimir Braverman, Dan Feldman, Harry Lang, Adiel Statman, and Samson Zhou. New frameworks for offline and streaming coreset constructions. arXiv preprint arXiv:1612.00889 , 2016

    arXiv 2016

  4. [4]

    V. Binh. Fraud eCommerce: Detecting Anomalous Transactions . https://www.kaggle.com/datasets/vbinh002/fraud-ecommerce, 2025. Kaggle dataset

    work page 2025

  5. [5]

    Data-dependent coresets for compressing neural networks with applications to generalization bounds

    Cenk Baykal, Lucas Liebenwein, Igor Gilitschenski, Dan Feldman, and Daniela Rus. Data-dependent coresets for compressing neural networks with applications to generalization bounds. In International Conference on Learning Representations

    work page

  6. [6]

    Improved coresets for euclidean k -means

    Vincent Cohen-Addad, Kasper Green Larsen, David Saulpic, Chris Schwiegelshohn, and Omar Ali Sheikh-Omar. Improved coresets for euclidean k -means. Advances in Neural Information Processing Systems , 35:2679--2694, 2022

    work page 2022

  7. [7]

    XGBoost : A scalable tree boosting system

    Tianqi Chen and Carlos Guestrin. XGBoost : A scalable tree boosting system. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , KDD '16, pages 785--794, New York, NY, USA, 2016. ACM

    work page 2016

  8. [8]

    Xgboost: A scalable tree boosting system

    Tianqi Chen and Carlos Guestrin. Xgboost: A scalable tree boosting system. In Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining , pages 785--794, 2016

    2016

Show all 42 references
  1. [9]

    Libsvm: A library for support vector machines

    Chih-Chung Chang and Chih-Jen Lin. Libsvm: A library for support vector machines. ACM transactions on intelligent systems and technology (TIST) , 2(3):1--27, 2011

    2011

  2. [10]

    Lp row sampling by lewis weights

    Michael B Cohen and Richard Peng. Lp row sampling by lewis weights. In Proceedings of the forty-seventh annual ACM symposium on Theory of computing , pages 183--192, 2015

    2015

  3. [11]

    Dataheroes: Automated framework for ml training set optimization and refinement, 2022

    DataHeroes. Dataheroes: Automated framework for ml training set optimization and refinement, 2022

    2022

  4. [12]

    Coreset-based neural network compression

    Abhimanyu Dubey, Moitreya Chatterjee, and Narendra Ahuja. Coreset-based neural network compression. In Proceedings of the European Conference on Computer Vision (ECCV) , pages 454--470, 2018

    2018

  5. [13]

    Calibrating probability with undersampling for unbalanced classification

    Andrea Dal Pozzolo, Olivier Caelen, Reid A Johnson, and Gianluca Bontempi. Calibrating probability with undersampling for unbalanced classification. In 2015 IEEE symposium series on computational intelligence , pages 159--166. IEEE, 2015

    2015

  6. [14]

    Additive logistic regression: a statistical view of boosting (with discussion and a rejoinder by the authors)

    Jerome Friedman, Trevor Hastie, and Robert Tibshirani. Additive logistic regression: a statistical view of boosting (with discussion and a rejoinder by the authors). The annals of statistics , 28(2):337--407, 2000

    2000

  7. [15]

    Greedy function approximation: a gradient boosting machine

    Jerome H Friedman. Greedy function approximation: a gradient boosting machine. Annals of statistics , pages 1189--1232, 2001

    2001

  8. [16]

    Turning big data into tiny data: Constant-size coresets for k-means, pca, and projective clustering

    Dan Feldman, Melanie Schmidt, and Christian Sohler. Turning big data into tiny data: Constant-size coresets for k-means, pca, and projective clustering. SIAM Journal on Computing , 49(3):601--657, 2020

    2020

  9. [17]

    Deepcore: A comprehensive library for coreset selection in deep learning

    Chengcheng Guo, Bo Zhao, and Yanbing Bai. Deepcore: A comprehensive library for coreset selection in deep learning. In International Conference on Database and Expert Systems Applications , pages 181--195. Springer, 2022

    2022

  10. [18]

    Harris, K

    Charles R. Harris, K. Jarrod Millman, Stéfan J van der Walt, Ralf Gommers, Pauli Virtanen, David Cournapeau, Eric Wieser, Julian Taylor, Sebastian Berg, Nathaniel J. Smith, Robert Kern, Matti Picus, Stephan Hoyer, Marten H. van Kerkwijk, Matthew Brett, Allan Haldane, Jaime Fer...

    2020

  11. [19]

    Ieee-cis fraud detection

    IEEE Computational Intelligence Society and Vesta Corporation. Ieee-cis fraud detection. https://www.kaggle.com/competitions/ieee-fraud-detection, October 2019. Accessed: 2025-06-20

    2019

  12. [20]

    Finite dimensional subspaces of lp

    William B Johnson and Gideon Schechtman. Finite dimensional subspaces of lp. Handbook of the geometry of Banach spaces , 1:837--870, 2001

    2001

  13. [21]

    Coresets for decision trees of signals

    Ibrahim Jubran, Ernesto Evgeniy Sanches Shayda, Ilan I Newman, and Dan Feldman. Coresets for decision trees of signals. Advances in Neural Information Processing Systems , 34:30352--30364, 2021

    2021

  14. [22]

    Sets clustering

    Ibrahim Jubran, Murad Tukan, Alaa Maalouf, and Dan Feldman. Sets clustering. In International Conference on Machine Learning , pages 4994--5005. PMLR, 2020

    2020

  15. [23]

    Umap: Uniform manifold approximation and projection for dimension reduction

    Leland McInnes, John Healy, and James Melville. Umap: Uniform manifold approximation and projection for dimension reduction. arXiv preprint arXiv:1802.03426 , 2018

    2018 arXiv

  16. [24]

    p-generalized probit regression and scalable maximum likelihood estimation via sketching and coresets

    Alexander Munteanu, Simon Omlor, and Christian Peters. p-generalized probit regression and scalable maximum likelihood estimation via sketching and coresets. In International Conference on Artificial Intelligence and Statistics , pages 2073--2100. PMLR, 2022

    2022

  17. [25]

    On coresets for logistic regression

    Alexander Munteanu, Chris Schwiegelshohn, Christian Sohler, and David Woodruff. On coresets for logistic regression. Advances in Neural Information Processing Systems , 31, 2018

    2018

  18. [26]

    Autocoreset: an automatic practical coreset construction framework

    Alaa Maalouf, Murad Tukan, Vladimir Braverman, and Daniela Rus. Autocoreset: an automatic practical coreset construction framework. In International Conference on Machine Learning , pages 23451--23466. PMLR, 2023

    2023

  19. [27]

    Coresets for data discretization and sine wave fitting

    Alaa Maalouf, Murad Tukan, Eric Price, Daniel M Kane, and Dan Feldman. Coresets for data discretization and sine wave fitting. In International Conference on Artificial Intelligence and Statistics , pages 10622--10639. PMLR, 2022

    2022

  20. [28]

    Early stopping-but when? In Neural Networks: Tricks of the trade , pages 55--69

    Lutz Prechelt. Early stopping-but when? In Neural Networks: Tricks of the trade , pages 55--69. Springer, 2002

    2002

  21. [29]

    Scikit-learn: Machine learning in python

    Fabian Pedregosa, Ga \"e l Varoquaux, Alexandre Gramfort, Vincent Michel, Bertrand Thirion, Olivier Grisel, Mathieu Blondel, Peter Prettenhofer, Ron Weiss, Vincent Dubourg, et al. Scikit-learn: Machine learning in python. Journal of machine learning research , 12(Oct):2825--2830, 2011

    2011

  22. [30]

    An efficient drifters deployment strategy to evaluate water current velocity fields

    Murad Tukan, Eli Biton, and Roee Diamant. An efficient drifters deployment strategy to evaluate water current velocity fields. IEEE Journal of Oceanic Engineering , 2024

    2024

  23. [31]

    On coresets for support vector machines

    Murad Tukan, Cenk Baykal, Dan Feldman, and Daniela Rus. On coresets for support vector machines. Theoretical Computer Science , 890:171--191, 2021

    2021

  24. [32]

    Generic coreset for scalable learning of monotonic kernels: Logistic regression, sigmoid and more

    Elad Tolochinksy, Ibrahim Jubran, and Dan Feldman. Generic coreset for scalable learning of monotonic kernels: Logistic regression, sigmoid and more. In International Conference on Machine Learning , pages 21520--21547. PMLR, 2022

    2022

  25. [33]

    Coresets for near-convex functions

    Murad Tukan, Alaa Maalouf, and Dan Feldman. Coresets for near-convex functions. Advances in Neural Information Processing Systems , 33:997--1009, 2020

    2020

  26. [34]

    Pruning neural networks via coresets and convex geometry: Towards no assumptions

    Murad Tukan, Loay Mualem, and Alaa Maalouf. Pruning neural networks via coresets and convex geometry: Towards no assumptions. Advances in Neural Information Processing Systems , 35:38003--38019, 2022

    2022

  27. [35]

    New coresets for projective clustering and applications

    Murad Tukan, Xuan Wu, Samson Zhou, Vladimir Braverman, and Dan Feldman. New coresets for projective clustering and applications. In International Conference on Artificial Intelligence and Statistics , pages 5391--5415. PMLR, 2022

    2022

  28. [36]

    Provable data subset selection for efficient neural networks training

    Murad Tukan, Samson Zhou, Alaa Maalouf, Daniela Rus, Vladimir Braverman, and Dan Feldman. Provable data subset selection for efficient neural networks training. In International Conference on Machine Learning , pages 34533--34555. PMLR, 2023

    2023

  29. [37]

    Oliphant, Matt Haberland, Tyler Reddy, David Cournapeau, Evgeni Burovski, Pearu Peterson, Warren Weckesser, Jonathan Bright, St \'e fan J

    Pauli Virtanen, Ralf Gommers, Travis E. Oliphant, Matt Haberland, Tyler Reddy, David Cournapeau, Evgeni Burovski, Pearu Peterson, Warren Weckesser, Jonathan Bright, St \'e fan J. van der Walt , Matthew Brett, Joshua Wilson, K. Jarrod Millman, Nikolay Mayorov, Andrew R. J. Nels...

    2020

  30. [38]

    Guido Van Rossum and Fred L. Drake. Python 3 Reference Manual . CreateSpace, Scotts Valley, CA, 2009

    2009

  31. [39]

    A near-linear algorithm for projective clustering integer points

    Kasturi Varadarajan and Xin Xiao. A near-linear algorithm for projective clustering integer points. In Proceedings of the twenty-third annual ACM-SIAM symposium on Discrete Algorithms , pages 1329--1342. SIAM, 2012

    2012

  32. [40]

    Daniel Whiteson. HEPMASS . UCI Machine Learning Repository, 2016. DOI : https://doi.org/10.24432/C5PP5W

    2016 doi

  33. [41]

    Automated filtering of human feedback data for aligning text-to-image diffusion models

    Yongjin Yang, Sihyeon Kim, Hojung Jung, Sangmin Bae, SangMook Kim, Se-Young Yun, and Kimin Lee. Automated filtering of human feedback data for aligning text-to-image diffusion models. In The Thirteenth International Conference on Learning Representations , 2025

    2025

  34. [42]

    Lima: Less is more for alignment

    Chunting Zhou, Pengfei Liu, Puxin Xu, Srinivasan Iyer, Jiao Sun, Yuning Mao, Xuezhe Ma, Avia Efrat, Ping Yu, Lili Yu, et al. Lima: Less is more for alignment. Advances in Neural Information Processing Systems , 36, 2024

    2024

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.