pith. machine review for the scientific record. sign in

arxiv: 2605.06278 · v1 · submitted 2026-05-07 · 💻 cs.LG · math.OC

Recognition: unknown

PACE: Prune-And-Compress Ensemble Models

Fabian Akkerman , Julien Ferry , Th\'eo Guyard , Thibaut Vidal
Authors on Pith no claims yet

Pith reviewed 2026-05-08 13:11 UTC · model grok-4.3

classification 💻 cs.LG math.OC
keywords ensemble pruningmodel compressionlearner diversityfaithfulness controlweak learnersprediction modelsmachine learning
0
0 comments X

The pith

PACE reduces ensemble model size by first generating diverse new learners then pruning the enriched set, with explicit control over faithfulness to the original.

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

Large ensembles of weak learners deliver strong predictions but create problems for deployment, interpretation, and verification tasks because of their bulk. PACE tackles this by running two phases in sequence: an active generation step that adds new learners according to a grounded procedure to increase diversity, followed by pruning that removes redundant members from the expanded collection. The framework lets users set how closely the reduced ensemble must match the original in its outputs. A sympathetic reader would care because this hybrid route claims to deliver smaller models than either pure pruning or pure compression can achieve on their own, while preserving performance and adding explicit faithfulness bounds. If the claim holds, practitioners gain a practical way to slim down ensembles without sacrificing the benefits that made them attractive in the first place.

Core claim

The paper claims that interleaving active learner generation to boost diversity with a subsequent pruning phase on the enriched ensemble produces smaller models that outperform prior pruning-only and compression-only methods while allowing principled, user-controlled faithfulness guarantees to the starting ensemble.

What carries the argument

PACE's two-phase strategy of active generation of new learners followed by pruning of the enriched ensemble, with tunable faithfulness constraints applied in both phases.

Load-bearing premise

The active generation step can reliably locate new learners that increase useful diversity without injecting bias or lowering overall quality, so that the later pruning step can safely shrink the model while keeping the promised performance and faithfulness.

What would settle it

Running PACE on a standard benchmark ensemble and finding that the final pruned version shows lower test accuracy than either the original ensemble or a strong pruning-only baseline, or that faithfulness metrics fall below the user-specified threshold.

Figures

Figures reproduced from arXiv: 2605.06278 by Fabian Akkerman, Julien Ferry, Th\'eo Guyard, Thibaut Vidal.

Figure 1
Figure 1. Figure 1: High-level description of PACE highlighting the active generation of improving learners (red) and the enforcement of faithfulness via separating sample generation (blue). The family of learners and the regions where faithfulness is enforced are tunable inputs of the method. 3 Building Improving Learners via Column Generation A key feature of PACE, which distinguishes it from previous approaches, is that it… view at source ↗
Figure 2
Figure 2. Figure 2: Time to achieve global faithfulness for Problem view at source ↗
Figure 3
Figure 3. Figure 3: Compressed ensemble size with varying parameters view at source ↗
Figure 4
Figure 4. Figure 4: Compressed ensemble size with varying parameters view at source ↗
Figure 5
Figure 5. Figure 5: Time to achieve global faithfulness for problem view at source ↗
Figure 6
Figure 6. Figure 6: Compressed ensemble size with varying parameters view at source ↗
Figure 7
Figure 7. Figure 7: Compressed ensemble size with varying parameters view at source ↗
Figure 8
Figure 8. Figure 8: Overall computational time (left) and number of separation problems solved (right) until view at source ↗
Figure 9
Figure 9. Figure 9: Overall computational time (left) and number of separation problems solved (right) until view at source ↗
Figure 10
Figure 10. Figure 10: Statistics of the improving learner generation phase of view at source ↗
Figure 11
Figure 11. Figure 11: Statistics of the improving learner generation phase of view at source ↗
read the original abstract

Ensemble models achieve state-of-the-art performance on prediction tasks, but usually require aggregating a large number of weak learners. This can hinder deployment, interpretability, and downstream tasks such as robustness verification. Remedies to this issue fall into two main camps: pruning, which discards redundant learners, and compression, which generates new ones from scratch. We introduce PACE, a framework that interleaves these paradigms in a two-phase strategy. First, new learners are actively generated via a theoretically grounded procedure to enhance the diversity of the initial ensemble. When no more relevant learners can be found, a second phase of pruning is performed on this enriched ensemble. During both operations, PACE allows fine control on the faithfulness to the original ensemble. Experiments show that our method outperforms prior pruning and compression methods while offering principled control of faithfulness guarantees.

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.

Referee Report

0 major / 2 minor

Summary. The paper introduces PACE, a two-phase framework for reducing the size of ensemble models. Phase one actively generates new learners via a theoretically grounded procedure to increase diversity in the initial ensemble. Phase two then performs pruning on the enriched ensemble. Both phases provide explicit control over faithfulness to the original ensemble. Experiments are reported to show outperformance relative to prior pruning and compression baselines.

Significance. If the active generation procedure is theoretically sound and the faithfulness controls are effective, the interleaving of generation and pruning could provide a principled method for producing smaller ensembles that retain predictive performance. This would be relevant for deployment constraints, interpretability, and downstream tasks such as robustness verification. The explicit faithfulness guarantees distinguish the approach from purely heuristic pruning or compression methods.

minor comments (2)
  1. The abstract refers to a 'theoretically grounded procedure' for learner generation and 'principled control of faithfulness guarantees,' but the provided text does not include the specific assumptions, theorems, or definitions that would allow verification of these claims.
  2. Experimental details (datasets, baselines, metrics, and statistical significance) are summarized at a high level; inclusion of concrete numbers, ablation studies, and reproducibility information would strengthen the presentation.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their review of our manuscript on PACE. We appreciate the accurate summary of the two-phase framework and the positive assessment of its potential significance for deployment, interpretability, and downstream tasks such as robustness verification. The conditional endorsement of the approach, contingent on the soundness of the active generation procedure and faithfulness controls, aligns with the core claims in the paper. Since no specific major comments were listed in the report, we have no point-by-point rebuttals to provide at this stage.

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The abstract and provided description outline a two-phase framework that first actively generates diverse learners via a theoretically grounded procedure and then applies pruning with faithfulness control. No equations, parameter-fitting steps, or self-referential definitions are visible that would make any claimed prediction or result equivalent to its inputs by construction. The approach is presented as interleaving existing pruning and compression paradigms with new elements, and the central claims rest on empirical experiments rather than reducing to fitted inputs or load-bearing self-citations. The derivation chain appears self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

No free parameters, axioms, or invented entities are specified in the abstract; the method is described at a high level without detailing any fitted quantities or new postulated constructs.

pith-pipeline@v0.9.0 · 5439 in / 1114 out tokens · 26386 ms · 2026-05-08T13:11:32.386396+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

68 extracted references · 1 canonical work pages

  1. [1]

    Boosting revisited: Benchmarking and advancing LP-based ensemble methods.Transactions on Machine Learning Research, 2025

    Fabian Akkerman, Julien Ferry, Christian Artigues, Emmanuel Hebrard, and Thibaut Vidal. Boosting revisited: Benchmarking and advancing LP-based ensemble methods.Transactions on Machine Learning Research, 2025

    2025

  2. [2]

    A novel pessimistic decision tree pruning approach for classification

    Abir Hossain Amee, Md Iftekhar Hossain, Sara Ferdous Khan, Dewan Md Farid, et al. A novel pessimistic decision tree pruning approach for classification. InInternational Conference on Electrical Information and Communication Technology (EICT), pages 1–6, 2023

    2023

  3. [3]

    Machine bias.ProPublica, 2016

    Julia Angwin, Jeff Larson, Surya Mattu, and Lauren Kirchner. Machine bias.ProPublica, 2016

    2016

  4. [4]

    Fast as CHITA: Neural network pruning with combinatorial optimization

    Riade Benbaki, Wenyu Chen, Xiang Meng, Hussein Hazimeh, Natalia Ponomareva, Zhe Zhao, and Rahul Mazumder. Fast as CHITA: Neural network pruning with combinatorial optimization. InInternational Conference on Machine Learning (ICML), pages 2031–2049, 2023

    2031

  5. [5]

    Bagging predictors.Machine learning, 24(2):123–140, 1996

    Leo Breiman. Bagging predictors.Machine learning, 24(2):123–140, 1996

    1996

  6. [6]

    Random forests.Machine learning, 45(1):5–32, 2001

    Leo Breiman. Random forests.Machine learning, 45(1):5–32, 2001

    2001

  7. [7]

    Chapman and Hall/CRC, 2017

    Leo Breiman, Jerome Friedman, Richard A Olshen, and Charles J Stone.Classification and regression trees. Chapman and Hall/CRC, 2017

    2017

  8. [8]

    Getting the most out of ensemble selection

    Rich Caruana, Art Munson, and Alexandru Niculescu-Mizil. Getting the most out of ensemble selection. InInternational Conference on Data Mining (ICDM), pages 828–833, 2006

    2006

  9. [9]

    Magorzata Charytanowicz, Jerzy Niewczas, Piotr Kulczycki, Piotr Kowalski, and Szymon Lukasik. Seeds. UCI Machine Learning Repository, 2010

    2010

  10. [10]

    Robustness verification of tree-based models

    Hongge Chen, Huan Zhang, Si Si, Yang Li, Duane Boning, and Cho-Jui Hsieh. Robustness verification of tree-based models. InAdvances in Neural Information Processing Systems (NeurIPS), pages 1–12, 2019

    2019

  11. [11]

    Hongrong Cheng, Miao Zhang, and Javen Qinfeng Shi. A survey on deep neural network pruning: Taxonomy, comparison, analysis, and recommendations.IEEE Transactions on Pattern Analysis and Machine Intelligence, 46(12):10558–10578, 2024

    2024

  12. [12]

    Boosting frank-wolfe by chasing gradients

    Cyrille Combettes and Sebastian Pokutta. Boosting frank-wolfe by chasing gradients. In International Conference on Machine Learning (ICML), pages 2111–2121, 2020

    2020

  13. [13]

    Dynamic decision tree ensembles for energy-efficient inference on IoT edge nodes.IEEE Internet of Things Journal, 11(1):742–757, 2024

    Francesco Daghero, Alessio Burrello, Enrico Macii, Paolo Montuschi, Massimo Poncino, and Daniele Jahier Pagliari. Dynamic decision tree ensembles for energy-efficient inference on IoT edge nodes.IEEE Internet of Things Journal, 11(1):742–757, 2024

    2024

  14. [14]

    Demiriz, K.P

    A. Demiriz, K.P. Bennett, and J. Shawe-Taylor. Linear programming boosting via column generation.Machine Learning, 46(1):225–254, 2002

    2002

  15. [15]

    Blossom: an anytime algorithm for computing optimal decision trees

    Emir Demirovi´c, Emmanuel Hebrard, and Louis Jean. Blossom: an anytime algorithm for computing optimal decision trees. InInternational Conference on Machine Learning (ICML), pages 7533–7562, 2023

    2023

  16. [16]

    Springer Science & Business Media, 2006

    Guy Desaulniers, Jacques Desrosiers, and Marius M Solomon.Column generation, volume 5. Springer Science & Business Media, 2006

    2006

  17. [17]

    Column generation

    Jacques Desrosiers, Marco Lübbecke, Guy Desaulniers, and Jean Bertrand Gauthier. Column generation. InBranch-and-Price, pages 43–102. Springer, 2026

    2026

  18. [18]

    Network pruning via transformable architecture search.Advances in Neural Information Processing Systems (NeurIPS), pages 1–12, 2019

    Xuanyi Dong and Yi Yang. Network pruning via transformable architecture search.Advances in Neural Information Processing Systems (NeurIPS), pages 1–12, 2019. 10

    2019

  19. [19]

    Free lunch in the forest: Functionally-identical pruning of boosted tree ensembles

    Youssouf Emine, Alexandre Forel, Idriss Malek, and Thibaut Vidal. Free lunch in the forest: Functionally-identical pruning of boosted tree ensembles. InAAAI Conference on Artificial Intelligence, pages 16488–16495, 2025

    2025

  20. [20]

    Trained random forests completely reveal your dataset

    Julien Ferry, Ricardo Fukasawa, Timothée Pascal, and Thibaut Vidal. Trained random forests completely reveal your dataset. InInternational Conference on Machine Learning (ICML), pages 13545–13569, 2024

    2024

  21. [21]

    Explainable ML challenge, 2025

    FICO. Explainable ML challenge, 2025. https://www.fico.com/en/newsroom/fico- expands-educational-analytics-challenge-program-three-new-historically-black-colleges-and- universities-educate-aspiring-data-scientists

    2025

  22. [22]

    Experiments with a new boosting algorithm

    Yoav Freund and Robert E Schapire. Experiments with a new boosting algorithm. InInterna- tional Conference on Machine Learning (ICML), pages 148–156, 1996

    1996

  23. [23]

    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

    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

  24. [24]

    Greedy function approximation: a gradient boosting machine.The annals of statistics, 29(5):1189–1232, 2001

    Jerome H Friedman. Greedy function approximation: a gradient boosting machine.The annals of statistics, 29(5):1189–1232, 2001

    2001

  25. [25]

    Extremely randomized trees.Machine learning, 63(1):3–42, 2006

    Pierre Geurts, Damien Ernst, and Louis Wehenkel. Extremely randomized trees.Machine learning, 63(1):3–42, 2006

    2006

  26. [26]

    Boosting in the limit: Maximizing the margin of learned ensembles

    Adam J Grove and Dale Schuurmans. Boosting in the limit: Maximizing the margin of learned ensembles. InAAAI Conference on Artificial Intelligence, pages 692–699, 1998

    1998

  27. [27]

    Gurobi Optimizer Reference Manual, 2026

    Gurobi Optimization, LLC. Gurobi Optimizer Reference Manual, 2026

    2026

  28. [28]

    PANDORA electronic collection

    Michael Harries.Splice-2 Comparative Evaluation: Electricity Pricing. PANDORA electronic collection. University of New South Wales, School of Computer Science and Engineering, 1999

    1999

  29. [29]

    Bootstrap.Wiley Interdisciplinary Reviews: Computational Statistics, 3(6): 497–526, 2011

    Tim Hesterberg. Bootstrap.Wiley Interdisciplinary Reviews: Computational Statistics, 3(6): 497–526, 2011

    2011

  30. [30]

    A survey on edge intelligence and lightweight machine learning support for future applications and services.J

    Kyle Hoffpauir, Jacob Simmons, Nikolas Schmidt, Rachitha Pittala, Isaac Briggs, Shanmukha Makani, and Yaser Jararweh. A survey on edge intelligence and lightweight machine learning support for future applications and services.J. Data and Information Quality, 15(2), 2023

    2023

  31. [31]

    Spambase

    Mark Hopkins, Erik Reeber, George Forman, and Jaap Suermondt. Spambase. UCI Machine Learning Repository, 1999

    1999

  32. [32]

    Implementing logical connectives in constraint programming.Artificial Intelligence, 174(16-17):1407–1429, 2010

    Christopher Jefferson, Neil CA Moore, Peter Nightingale, and Karen E Petrie. Implementing logical connectives in constraint programming.Artificial Intelligence, 174(16-17):1407–1429, 2010

    2010

  33. [33]

    Learning optimal fair decision trees: Trade-offs between interpretability, fairness, and accuracy

    Nathanael Jo, Sina Aghaei, Jack Benson, Andres Gomez, and Phebe Vayanos. Learning optimal fair decision trees: Trade-offs between interpretability, fairness, and accuracy. InAAAI/ACM Conference on AI, Ethics, and Society, pages 181–192, 2023

    2023

  34. [34]

    Resource-efficient machine learning in 2 KB RAM for the internet of things

    Ashish Kumar, Saurabh Goyal, and Manik Varma. Resource-efficient machine learning in 2 KB RAM for the internet of things. InInternational Conference on Machine Learning (ICML), pages 1935–1944, 2017

    1935

  35. [35]

    Pruning vs quantization: Which is better?Advances in Neural Information Processing Systems (NeurIPS), pages 62414–62427, 2023

    Andrey Kuzmin, Markus Nagel, Mart Van Baalen, Arash Behboodi, and Tijmen Blankevoort. Pruning vs quantization: Which is better?Advances in Neural Information Processing Systems (NeurIPS), pages 62414–62427, 2023

    2023

  36. [36]

    Diversity regularized ensemble pruning

    Nan Li, Yang Yu, and Zhi-Hua Zhou. Diversity regularized ensemble pruning. InJoint European conference on machine learning and knowledge discovery in databases (ECML-PKDD), pages 330–345, 2012. 11

    2012

  37. [37]

    Lost in pruning: The effects of pruning neural networks beyond test accuracy

    Lucas Liebenwein, Cenk Baykal, Brandon Carter, David Gifford, and Daniela Rus. Lost in pruning: The effects of pruning neural networks beyond test accuracy. InAnnual Conference on Machine Learning and Systems (MLSys), pages 93–138, 2021

    2021

  38. [38]

    Forestprune: Compact depth-pruned tree ensembles

    Brian Liu and Rahul Mazumder. Forestprune: Compact depth-pruned tree ensembles. In International Conference on Artificial Intelligence and Statistics (AISTATS), pages 9417–9428, 2023

    2023

  39. [39]

    Isolation forest

    Fei Tony Liu, Kai Ming Ting, and Zhi-Hua Zhou. Isolation forest. InInternational Conference on Data Mining (ICDM), pages 413–422, 2008

    2008

  40. [40]

    Isolation-based anomaly detection.ACM Transactions on Knowledge Discovery from Data, 6(1):1–39, 2012

    Fei Tony Liu, Kai Ming Ting, and Zhi-Hua Zhou. Isolation-based anomaly detection.ACM Transactions on Knowledge Discovery from Data, 6(1):1–39, 2012

    2012

  41. [41]

    Rethinking the value of network pruning

    Zhuang Liu, Mingjie Sun, Tinghui Zhou, Gao Huang, and Trevor Darrell. Rethinking the value of network pruning. InInternational Conference on Learning Representations (ICLR), 2019

    2019

  42. [42]

    Semi-infinite programming.European journal of operational research, 180(2):491–518, 2007

    Marco López and Georg Still. Semi-infinite programming.European journal of operational research, 180(2):491–518, 2007

    2007

  43. [43]

    Robert Lyon. HTRU2. UCI Machine Learning Repository, 2015

    2015

  44. [44]

    M5 accuracy competi- tion: Results, findings, and conclusions.International Journal of Forecasting, 38(4):1346–1364,

    Spyros Makridakis, Evangelos Spiliotis, and Vassilios Assimakopoulos. M5 accuracy competi- tion: Results, findings, and conclusions.International Journal of Forecasting, 38(4):1346–1364,

  45. [45]

    Special Issue: M5 competition

  46. [46]

    Using boosting to prune bagging ensembles

    Gonzalo Martinez-Munoz and Alberto Suárez. Using boosting to prune bagging ensembles. Pattern Recognition Letters, 28(1):156–165, 2007

    2007

  47. [47]

    An analysis of ensemble pruning techniques based on ordered aggregation.IEEE Transactions on Pattern Analysis and Machine Intelligence, 31(2):245–259, 2008

    Gonzalo Martinez-Munoz, Daniel Hernández-Lobato, and Alberto Suárez. An analysis of ensemble pruning techniques based on ordered aggregation.IEEE Transactions on Pattern Analysis and Machine Intelligence, 31(2):245–259, 2008

    2008

  48. [48]

    Falcon: Flop-aware com- binatorial optimization for neural network pruning

    Xiang Meng, Wenyu Chen, Riade Benbaki, and Rahul Mazumder. Falcon: Flop-aware com- binatorial optimization for neural network pruning. InInternational Conference on Artificial Intelligence and Statistics (AISTATS), pages 4384–4392, 2024

    2024

  49. [49]

    Optimal counterfactual explanations in tree ensembles

    Axel Parmentier and Thibaut Vidal. Optimal counterfactual explanations in tree ensembles. In International Conference on Machine Learning (ICML), pages 8422–8431, 2021

    2021

  50. [50]

    Pruning an ensemble of classifiers via reinforcement learning.Neurocomputing, 72(7-9):1900–1909, 2009

    Ioannis Partalas, Grigorios Tsoumakas, and Ioannis Vlahavas. Pruning an ensemble of classifiers via reinforcement learning.Neurocomputing, 72(7-9):1900–1909, 2009

    1900

  51. [51]

    Scikit-learn: Machine learning in Python.Journal of Machine Learning Research, 12:2825–2830, 2011

    Fabian Pedregosa, Gaël Varoquaux, Alexandre Gramfort, Vincent Michel, Bertrand Thirion, Olivier Grisel, Mathieu Blondel, Peter Prettenhofer, Ron Weiss, Vincent Dubourg, Jake Vander- plas, Alexandre Passos, David Cournapeau, Matthieu Brucher, Matthieu Perrot, and Édouard Duchesnay. Scikit-learn: Machine learning in Python.Journal of Machine Learning Resear...

    2011

  52. [52]

    CP-SAT, 2024

    Laurent Perron and Frédéric Didier. CP-SAT, 2024

    2024

  53. [53]

    Pareto ensemble pruning

    Chao Qian, Yang Yu, and Zhi-Hua Zhou. Pareto ensemble pruning. InAAAI Conference on Artificial Intelligence, pages 2935–2941, 2015

    2015

  54. [54]

    Random forest.Journal of insurance medicine, 47(1):31–39, 2017

    Steven J Rigatti. Random forest.Journal of insurance medicine, 47(1):31–39, 2017

    2017

  55. [55]

    Elsevier, 2006

    Francesca Rossi, Peter Van Beek, and Toby Walsh.Handbook of constraint programming. Elsevier, 2006

    2006

  56. [56]

    Ensemble learning: A survey.Wiley interdisciplinary reviews: data mining and knowledge discovery, 8(4):e1249, 2018

    Omer Sagi and Lior Rokach. Ensemble learning: A survey.Wiley interdisciplinary reviews: data mining and knowledge discovery, 8(4):e1249, 2018

    2018

  57. [57]

    Explaining adaboost

    Robert E Schapire. Explaining adaboost. InEmpirical inference: festschrift in honor of vladimir N. Vapnik, pages 37–52. Springer Berlin Heidelberg, 2013. 12

    2013

  58. [58]

    On the dual formulation of boosting algorithms.IEEE Transactions on Pattern Analysis and Machine Intelligence, 32(12):2216–2231, 2010

    Chunhua Shen and Hanxi Li. On the dual formulation of boosting algorithms.IEEE Transactions on Pattern Analysis and Machine Intelligence, 32(12):2216–2231, 2010

    2010

  59. [59]

    Classification of radar returns from the ionosphere using neural networks.Johns Hopkins APL Technical Digest, 10(3): 262–266, 1989

    Vincent G Sigillito, Simon P Wing, Larrie V Hutton, and Kile B Baker. Classification of radar returns from the ionosphere using neural networks.Johns Hopkins APL Technical Digest, 10(3): 262–266, 1989

    1989

  60. [60]

    Pima Indians diabetes data set, 1990

    Peter Turney. Pima Indians diabetes data set, 1990

    1990

  61. [61]

    van Rijn, Bernd Bischl, and Luis Torgo

    Joaquin Vanschoren, Jan N. van Rijn, Bernd Bischl, and Luis Torgo. OpenML: Networked science in machine learning.ACM SIGKDD Explorations Newsletter, 15(2):49–60, 2014

    2014

  62. [62]

    Born-again tree ensembles

    Thibaut Vidal and Maximilian Schiffer. Born-again tree ensembles. InInternational Conference on Machine Learning (ICML), pages 9743–9753, 2020

    2020

  63. [63]

    Entropy regularized lpboost

    Manfred K Warmuth, Karen A Glocer, and SVN Vishwanathan. Entropy regularized lpboost. InInternational Conference on Algorithmic Learning Theory (ALT), pages 256–271, 2008

    2008

  64. [64]

    Weiss and Nitin Indurkhya

    Sholom M. Weiss and Nitin Indurkhya. Rule-based machine learning methods for functional prediction, 1995. ArXiv preprint

    1995

  65. [65]

    Ensemble pruning via semi-definite programming.Journal of machine learning research, 7(7), 2006

    Yi Zhang, Samuel Burer, W Nick Street, Kristin P Bennett, and Emilio Parrado-Hernández. Ensemble pruning via semi-definite programming.Journal of machine learning research, 7(7), 2006

    2006

  66. [66]

    Advancing model pruning via bi-level optimization.Advances in Neural Information Processing Systems (NeurIPS), pages 18309–18326, 2022

    Yihua Zhang, Yuguang Yao, Parikshit Ram, Pu Zhao, Tianlong Chen, Mingyi Hong, Yanzhi Wang, and Sijia Liu. Advancing model pruning via bi-level optimization.Advances in Neural Information Processing Systems (NeurIPS), pages 18309–18326, 2022

    2022

  67. [67]

    Interpreting models via single tree approximation.arXiv preprint arXiv:1610.09036, 2016

    Yichen Zhou and Giles Hooker. Interpreting models via single tree approximation.arXiv preprint arXiv:1610.09036, 2016

    work page arXiv 2016

  68. [68]

    Breast Cancer

    Matjaz Zwitter and Milan Soklic. Breast Cancer. UCI Machine Learning Repository, 1988. 13 A Related Works This appendix provides a review of related works on ensemble training, pruning, and compression techniques, as well as connections to other machine learning models beyond ensemble models. A.1 Ensemble Model Training Training an ensemble model amounts ...

    1988