pith. sign in

arxiv: 2606.05380 · v1 · pith:3LITQGGVnew · submitted 2026-06-03 · 💻 cs.DS · cs.LG

Learning-Augmented Online Minimization with Dual Predictions

Christian Coester , Alexa Tudose , Alexander Turoczy This is my paper

Pith reviewed 2026-06-28 03:21 UTC · model grok-4.3

classification 💻 cs.DS cs.LG
keywords learning-augmented algorithmsonline minimizationdual predictionsmetrical task systemslaminar set covercompetitive analysisk-server problemparking permit problem
0
0 comments X

The pith

Machine-learned predictions of dual linear program solutions improve competitive ratios for online minimization problems including metrical task systems and laminar set cover.

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

The paper presents algorithms for two classes of online minimization problems that incorporate predictions of optimal dual linear program solutions. These dual predictions remain stable under small perturbations to the input, unlike primal solutions, so they can be learned reliably from families of similar instances. The resulting algorithms obtain improved theoretical performance bounds compared with classical online methods. Experiments on the k-server problem and the parking permit problem provide empirical support. A reader would care because the work shows how learning can be integrated into online decision-making when only stable aspects of the optimum are predictable.

Core claim

Predicting an optimal solution to the dual of the linear programming relaxation allows the design of learning-augmented algorithms that achieve better competitive ratios for metrical task systems and laminar set cover. Because dual solutions change little when the instance is perturbed slightly, consistent and learnable predictions exist for related problem families, and this approach constitutes the first effective use of dual predictions in online minimization.

What carries the argument

Machine-learned predictions of optimal solutions to the dual linear program, which supply stable guidance that the online algorithm can follow to reduce cost.

If this is right

  • Metrical task systems obtain improved competitive ratios when dual predictions are available.
  • Laminar set cover obtains improved competitive ratios when dual predictions are available.
  • The k-server problem exhibits better empirical performance with dual predictions than without them.
  • The parking permit problem exhibits better empirical performance with dual predictions than without them.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same dual-prediction technique could extend to other online minimization problems whose LP relaxations have stable dual optima.
  • If dual stability holds across broader instance distributions, learning-augmented methods may become viable for additional online problems where primal optima fluctuate.
  • Empirical tests on real-world sequences of similar instances would measure how accurately dual predictions can be learned in practice.

Load-bearing premise

Dual solutions to the linear program remain stable under tiny changes to the input instance.

What would settle it

A concrete family of similar instances in which the optimal dual solutions differ enough that no single prediction yields better performance than the non-learning baseline.

Figures

Figures reproduced from arXiv: 2606.05380 by Alexander Turoczy, Alexa Tudose, Christian Coester.

Figure 1
Figure 1. Figure 1: Experiments on the Parking Permit Problem. Shades denote approximate 95% confidence intervals. the factor by which costs of subsequent permits decrease. We compare the performance of our learning-augmented algorithm (with α = 0.5) relative to state-of-the-art O(K)- competitive deterministic and O(log K)-competitive ran￾domized classical online algorithms for PPP. We use the randomized algorithm as the inpu… view at source ↗
Figure 2
Figure 2. Figure 2: The average competitive ratio of each algorithm, where the number of points in the metric space is fixed to 10, and the number of servers varies from k = 2 to k = 9. Shades denote approximate 95% confidence intervals. ables corresponding to the 1:00–1:15 time block across all training instances. This choice of prediction reflects the idea that request patterns remain relatively stable within short time int… view at source ↗
read the original abstract

We present learning-augmented algorithms for two general classes of online minimization problems: metrical task systems and laminar set cover. Both algorithms achieve improved theoretical guarantees using machine-learned predictions of an optimal solution to the dual linear program. Unlike optimal primal solutions, which can change drastically under tiny instance perturbations, these dual solutions are much more stable, which ensures the existence of good (and learnable) predictions for families of similar instances. While previous work has used dual predictions in offline settings and for online maximization problems, our algorithms are, to the best of our knowledge, the first demonstration that such dual predictions can be effective for online minimization. Our theoretical results are complemented by experiments on the $k$-server problem and the parking permit problem.

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

2 major / 1 minor

Summary. The paper presents learning-augmented algorithms for metrical task systems and laminar set cover that use machine-learned predictions of optimal dual LP solutions to obtain improved competitive ratios for online minimization. It asserts that dual solutions are much more stable than primal solutions under small instance perturbations, which enables the existence of good, learnable predictions, and positions this as the first such demonstration for minimization problems. Theoretical claims are complemented by experiments on the k-server and parking permit problems.

Significance. If the stability assertion can be backed by formal bounds, the approach would provide a new technique for learning-augmented online minimization, potentially improving upon existing primal-prediction methods in settings where dual variables exhibit greater robustness.

major comments (2)
  1. [Abstract] Abstract: the central claim that dual solutions are 'much more stable' than primal ones under tiny perturbations (which 'ensures the existence of good (and learnable) predictions') is asserted without any quantitative stability bound, Lipschitz constant on the dual map, or lemma establishing this property for metrical task systems or laminar set cover.
  2. [Abstract] Abstract / theoretical results: no reduction is supplied showing that the asserted dual stability implies PAC-learnability of the predictor or directly yields the claimed competitive-ratio improvements; the improved guarantees could therefore be artifacts of the specific prediction model rather than a general consequence of stability.
minor comments (1)
  1. [Abstract] The abstract would be strengthened by stating the concrete competitive-ratio improvements obtained by the new algorithms.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive feedback on our manuscript. We address the two major comments point by point below, indicating planned revisions where appropriate.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central claim that dual solutions are 'much more stable' than primal ones under tiny perturbations (which 'ensures the existence of good (and learnable) predictions') is asserted without any quantitative stability bound, Lipschitz constant on the dual map, or lemma establishing this property for metrical task systems or laminar set cover.

    Authors: We agree that the abstract makes a qualitative assertion about dual stability without supplying a formal Lipschitz bound or dedicated lemma. The full paper motivates this via the structure of the dual LPs (potentials in metrical task systems and multipliers in laminar set cover), where small cost perturbations induce only gradual changes in optimal dual values, in contrast to primal solutions that may switch bases abruptly. No general quantitative bound is derived because the modulus of continuity depends on instance-specific parameters such as the metric diameter or laminar depth. We will revise the abstract to remove the parenthetical claim about ensuring learnability and replace it with a more measured statement that dual predictions are empirically robust for families of similar instances, with supporting discussion and examples added to the introduction. revision: yes

  2. Referee: [Abstract] Abstract / theoretical results: no reduction is supplied showing that the asserted dual stability implies PAC-learnability of the predictor or directly yields the claimed competitive-ratio improvements; the improved guarantees could therefore be artifacts of the specific prediction model rather than a general consequence of stability.

    Authors: The competitive-ratio improvements are obtained directly from the analysis (Theorems 3 and 5), which bound the online cost by the optimal cost plus a term linear in the dual prediction error; the proofs do not rely on an external stability-to-learnability reduction. The stability claim is used only to argue that good predictors exist and can be trained on historical data from similar instances. While we do not supply an explicit PAC-learnability reduction or a general theorem linking arbitrary stability to learnability, the specific dual maps arising in these problems are Lipschitz-continuous on compact instance sets, permitting standard regression techniques. We will add a short clarifying paragraph after the main theorems explaining that the error bounds hold for any predictor (learned or otherwise) and that stability merely justifies why low-error predictors are attainable in practice. revision: yes

Circularity Check

0 steps flagged

No circularity: central stability claim is an assertion without reduction to fitted inputs or self-citations

full rationale

The abstract asserts that dual LP solutions are 'much more stable' than primal ones under perturbations, 'which ensures the existence of good (and learnable) predictions,' but supplies no equations, fitted parameters, or self-citations that would make this claim reduce to its own inputs by construction. No derivation chain is visible that renames a known result, smuggles an ansatz via citation, or calls a fitted quantity a 'prediction.' The paper's claim of being the 'first demonstration' for online minimization is a novelty statement, not a load-bearing self-citation. Per the hard rules, absent any quotable reduction (e.g., Eq. X defined in terms of Y), the finding is no significant circularity.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review; no explicit free parameters, axioms, or invented entities are stated. Standard online-algorithm assumptions (adversarial input, competitive analysis) are implicit but not detailed.

pith-pipeline@v0.9.1-grok · 5649 in / 1050 out tokens · 33308 ms · 2026-06-28T03:21:16.966946+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

94 extracted references · 10 canonical work pages

  1. [1]

    Priyank Agrawal and Eric Balkanski and Vasilis Gkatzelis and Tingting Ou and Xizhi Tan , title =. Math. Oper. Res. , volume =. 2024 , url =

    2024

  2. [2]

    2009 , url =

    Noga Alon and Baruch Awerbuch and Yossi Azar and Niv Buchbinder and Joseph Naor , title =. 2009 , url =

    2009

  3. [3]

    Learning-Augmented Online Covering Problems , journal =

    Afrouz Jabal Ameli and Laura Sanit. Learning-Augmented Online Covering Problems , journal =. 2025 , url =

    2025

  4. [4]

    arXiv preprint arXiv:2402.13530 , year=

    Best of many in both worlds: Online resource allocation with predictions under unknown arrival model , author=. arXiv preprint arXiv:2402.13530 , year=

    arXiv

  5. [5]

    International Conference on Machine Learning,

    Keerti Anand and Rong Ge and Amit Kumar and Debmalya Panigrahi , title =. International Conference on Machine Learning,. 2022 , url =

    2022

  6. [6]

    Online computation with untrusted advice , journal =

    Spyros Angelopoulos and Christoph D. Online computation with untrusted advice , journal =. 2024 , url =

    2024

  7. [7]

    Paging with Succinct Predictions , booktitle =

    Antonios Antoniadis and Joan Boyar and Marek Eli. Paging with Succinct Predictions , booktitle =. 2023 , url =

    2023

  8. [8]

    Learning-Augmented Dynamic Power Management with Multiple States via New Ski Rental Bounds , booktitle =

    Antonios Antoniadis and Christian Coester and Marek Eli. Learning-Augmented Dynamic Power Management with Multiple States via New Ski Rental Bounds , booktitle =. 2021 , url =

    2021

  9. [9]

    Online Metric Algorithms with Untrusted Predictions , journal =

    Antonios Antoniadis and Christian Coester and Marek Eli. Online Metric Algorithms with Untrusted Predictions , journal =. 2023 , url =

    2023

  10. [10]

    Mixing Predictions for Online Metric Algorithms , booktitle =

    Antonios Antoniadis and Christian Coester and Marek Eli. Mixing Predictions for Online Metric Algorithms , booktitle =. 2023 , url =

    2023

  11. [11]

    Approximation algorithms for combinatorial optimization with predictions , booktitle =

    Antonios Antoniadis and Marek Eli. Approximation algorithms for combinatorial optimization with predictions , booktitle =. 2025 , url =

    2025

  12. [12]

    53rd Annual

    Yossi Azar and Stefano Leonardi and Noam Touitou , title =. 53rd Annual. 2021 , url =

    2021

  13. [13]

    Proceedings of the 2022

    Yossi Azar and Debmalya Panigrahi and Noam Touitou , title =. Proceedings of the 2022. 2022 , url =

    2022

  14. [14]

    Advances in Neural Information Processing Systems 36, NeurIPS , year =

    Yossi Azar and Debmalya Panigrahi and Noam Touitou , title =. Advances in Neural Information Processing Systems 36, NeurIPS , year =

  15. [15]

    Advances in Neural Information Processing Systems 36, NeurIPS , year =

    Xingjian Bai and Christian Coester , title =. Advances in Neural Information Processing Systems 36, NeurIPS , year =

  16. [16]

    Advances in Neural Information Processing Systems 38, NeurIPS , year =

    Eric Balkanski and Vasilis Gkatzelis and Golnoosh Shahkarami , title =. Advances in Neural Information Processing Systems 38, NeurIPS , year =

  17. [17]

    Energy-Efficient Scheduling with Predictions , booktitle =

    Eric Balkanski and No. Energy-Efficient Scheduling with Predictions , booktitle =. 2023 , url =

    2023

  18. [18]

    The Primal-Dual method for Learning Augmented Algorithms , booktitle =

  19. [19]

    Forty-first International Conference on Machine Learning,

    Evripidis Bampis and Bruno Escoffier and Michalis Xefteris , title =. Forty-first International Conference on Machine Learning,. 2024 , url =

    2024

  20. [20]

    Proceedings of the 2022

    Nikhil Bansal and Christian Coester and Ravi Kumar and Manish Purohit and Erik Vee , title =. Proceedings of the 2022. 2022 , url =

    2022

  21. [21]

    Yair Bartal and Elias Koutsoupias , title =. Theor. Comput. Sci. , volume =. 2004 , url =

    2004

  22. [22]

    Advances in Neural Information Processing Systems 38, NeurIPS , year =

    Ziyad Benomar and Christian Coester , title =. Advances in Neural Information Processing Systems 38, NeurIPS , year =

  23. [23]

    Forty-first International Conference on Machine Learning,

    Ziyad Benomar and Vianney Perchet , title =. Forty-first International Conference on Machine Learning,. 2024 , url =

    2024

  24. [24]

    Avrim Blum and Carl Burch , title =. Mach. Learn. , volume =. 2000 , url =

    2000

  25. [25]

    Saks , title =

    Allan Borodin and Nathan Linial and Michael E. Saks , title =. J. 1992 , url =

    1992

  26. [26]

    The Randomized k-Server Conjecture Is False! , booktitle =

    S. The Randomized k-Server Conjecture Is False! , booktitle =. 2023 , url =

    2023

  27. [27]

    Metrical Task Systems on Trees via Mirror Descent and Unfair Gluing , journal =

    S. Metrical Task Systems on Trees via Mirror Descent and Unfair Gluing , journal =. 2021 , url =

    2021

  28. [28]

    Niv Buchbinder and Joseph Naor , title =. Found. Trends Theor. Comput. Sci. , volume =. 2009 , url =

    2009

  29. [29]

    Chen and Sandeep Silwal and Ali Vakilian and Fred Zhang , title =

    Justin Y. Chen and Sandeep Silwal and Ali Vakilian and Fred Zhang , title =. International Conference on Machine Learning,. 2022 , url =

    2022

  30. [30]

    Conference on Learning Theory,

    Nicolas Christianson and Tinashe Handina and Adam Wierman , title =. Conference on Learning Theory,. 2022 , url =

    2022

  31. [31]

    International Conference on Artificial Intelligence and Statistics, AISTATS , pages =

    Nicolas Christianson and Junxuan Shen and Adam Wierman , title =. International Conference on Artificial Intelligence and Statistics, AISTATS , pages =. 2023 , url =

    2023

  32. [32]

    Karloff and T

    Marek Chrobak and Howard J. Karloff and T. H. Payne and Sundar Vishwanathan , title =. Proceedings of the First Annual. 1990 , url =

    1990

  33. [33]

    Larmore , title =

    Marek Chrobak and Lawrence L. Larmore , title =. On-Line Algorithms, Proceedings of a. 1991 , url =

    1991

  34. [34]

    Lee , title =

    Christian Coester and James R. Lee , title =. Theory Comput. , volume =. 2022 , url =

    2022

  35. [35]

    53rd International Colloquium on Automata, Languages, and Programming,

    Christian Coester and Alexa Tudose , title =. 53rd International Colloquium on Automata, Languages, and Programming,

  36. [36]

    50th International Colloquium on Automata, Languages, and Programming,

    Ilan Reuven Cohen and Debmalya Panigrahi , title =. 50th International Colloquium on Automata, Languages, and Programming,. 2023 , url =. doi:10.4230/LIPICS.ICALP.2023.43 , timestamp =

    work page doi:10.4230/lipics.icalp.2023.43 2023

  37. [37]

    Advances in Neural Information Processing Systems 39, NeurIPS , year =

    Ilan Reuven Cohen and Debmalya Panigrahi , title =. Advances in Neural Information Processing Systems 39, NeurIPS , year =

  38. [38]

    Learning-Augmented Algorithms for

    Matei Gabriel Cosa and Marek Eli. Learning-Augmented Algorithms for. Forty-second International Conference on Machine Learning,. 2025 , url =

    2025

  39. [39]

    International Conference on Machine Learning,

    Sami Davies and Benjamin Moseley and Sergei Vassilvitskii and Yuyan Wang , title =. International Conference on Machine Learning,. 2023 , url =

    2023

  40. [40]

    Devanur and Thomas P

    Nikhil R. Devanur and Thomas P. Hayes , title =. Proceedings 10th. 2009 , url =

    2009

  41. [41]

    Proceedings of the 38th International Conference on Machine Learning,

    Ilias Diakonikolas and Vasilis Kontonis and Christos Tzamos and Ali Vakilian and Nikos Zarifis , title =. Proceedings of the 38th International Conference on Machine Learning,. 2021 , url =

    2021

  42. [42]

    Advances in Neural Information Processing Systems 34, NeurIPS , pages =

    Michael Dinitz and Sungjin Im and Thomas Lavastida and Benjamin Moseley and Sergei Vassilvitskii , title =. Advances in Neural Information Processing Systems 34, NeurIPS , pages =

  43. [43]

    Ergun and Zhili Feng and Sandeep Silwal and David P

    Jon C. Ergun and Zhili Feng and Sandeep Silwal and David P. Woodruff and Samson Zhou , title =. The Tenth International Conference on Learning Representations,. 2022 , url =

    2022

  44. [44]

    Mirrokni and Clifford Stein , title =

    Jon Feldman and Monika Henzinger and Nitish Korula and Vahab S. Mirrokni and Clifford Stein , title =. Algorithms -. 2010 , url =

    2010

  45. [45]

    Foster and Howard J

    Amos Fiat and Dean P. Foster and Howard J. Karloff and Yuval Rabani and Yiftach Ravid and Sundar Vishwanathan , title =. 1998 , url =

    1998

  46. [46]

    Vasilis Gkatzelis and Kostas Kollias and Alkmini Sgouritsa and Xizhi Tan , title =. 23rd. 2022 , url =

    2022

  47. [47]

    Proceedings of the 36th International Conference on Machine Learning,

    Sreenivas Gollapudi and Debmalya Panigrahi , title =. Proceedings of the 36th International Conference on Machine Learning,. 2019 , url =

    2019

  48. [48]

    Learning-Augmented Algorithms for Online Linear and Semidefinite Programming , booktitle =

    Elena Grigorescu and Young. Learning-Augmented Algorithms for Online Linear and Semidefinite Programming , booktitle =. 2022 , url =

    2022

  49. [49]

    Advances in Neural Information Processing Systems 35, NeurIPS , year =

    Anupam Gupta and Debmalya Panigrahi and Bernardo Subercaseaux and Kevin Sun , title =. Advances in Neural Information Processing Systems 35, NeurIPS , year =

  50. [50]

    Learning-Based Frequency Estimation Algorithms , booktitle =

    Chen. Learning-Based Frequency Estimation Algorithms , booktitle =. 2019 , url =

    2019

  51. [51]

    2022 , url =

    Zhihao Jiang and Debmalya Panigrahi and Kevin Sun , title =. 2022 , url =

    2022

  52. [52]

    Learning Predictions for Algorithms with Predictions , booktitle =

    Misha Khodak and Maria. Learning Predictions for Algorithms with Predictions , booktitle =. 2022 , url =

    2022

  53. [53]

    Proceedings of the 2020

    Silvio Lattanzi and Thomas Lavastida and Benjamin Moseley and Sergei Vassilvitskii , title =. Proceedings of the 2020. 2020 , url =

    2020

  54. [54]

    International Conference on Machine Learning,

    Silvio Lattanzi and Ola Svensson and Sergei Vassilvitskii , title =. International Conference on Machine Learning,. 2023 , url =

    2023

  55. [55]

    Ravi and Chenyang Xu , title =

    Thomas Lavastida and Benjamin Moseley and R. Ravi and Chenyang Xu , title =. 29th Annual European Symposium on Algorithms,. 2021 , url =

    2021

  56. [56]

    Shenoy , title =

    Adam Lechowicz and Nicolas Christianson and Bo Sun and Noman Bashir and Mohammad Hajiesmaili and Adam Wierman and Prashant J. Shenoy , title =. Forty-first International Conference on Machine Learning,. 2024 , url =

    2024

  57. [57]

    Proceedings of the 38th International Conference on Machine Learning,

    Shi Li and Jiayi Xian , title =. Proceedings of the 38th International Conference on Machine Learning,

  58. [58]

    Woodruff , title =

    Honghao Lin and Tian Luo and David P. Woodruff , title =. International Conference on Machine Learning,

  59. [59]

    Algorithms with predictions , author=

  60. [60]

    Alexander Lindermayr and Nicole Megow , title =

  61. [61]

    Algorithmica , volume =

    Alexander Lindermayr and Nicole Megow and Bertrand Simon , title =. Algorithmica , volume =

  62. [62]

    Thodoris Lykouris and Sergei Vassilvitskii , title =. J

  63. [63]

    46th Annual

    Adam Meyerson , title =. 46th Annual

  64. [64]

    Algorithms with Predictions , booktitle =

    Michael Mitzenmacher and Sergei Vassilvitskii , editor =. Algorithms with Predictions , booktitle =

  65. [65]

    Westbrook , editor =

    Steven Phillips and Jeffery R. Westbrook , editor =. On-line Algorithms , booktitle =

  66. [66]

    Advances in Neural Information Processing Systems 31 (NeurIPS) , pages =

    Manish Purohit and Zoya Svitkina and Ravi Kumar , title =. Advances in Neural Information Processing Systems 31 (NeurIPS) , pages =

  67. [67]

    Proceedings of the 2020

    Dhruv Rohatgi , title =. Proceedings of the 2020

    2020

  68. [68]

    Algorithms for Caching and

    Karim Abdel Sadek and Marek Eli. Algorithms for Caching and. The Twelfth International Conference on Learning Representations,

  69. [69]

    Daniel Dominic Sleator and Robert Endre Tarjan , title =. Commun. 1985 , url =

    1985

  70. [70]

    Advances in Neural Information Processing Systems 33, NeurIPS , year =

    Shufan Wang and Jian Li and Shiqiang Wang , title =. Advances in Neural Information Processing Systems 33, NeurIPS , year =

  71. [71]

    Approximation, Randomization, and Combinatorial Optimization

    Alexander Wei , title =. Approximation, Randomization, and Combinatorial Optimization. Algorithms and Techniques,

  72. [72]

    Advances in Neural Information Processing Systems 33 , year =

    Alexander Wei and Fred Zhang , title =. Advances in Neural Information Processing Systems 33 , year =

  73. [73]

    Mathematical Optimization Theory and Operations Research: 23rd International Conference, MOTOR , pages =

    Yaroslav, Kharchenko and Alexander, Kononov , title =. Mathematical Optimization Theory and Operations Research: 23rd International Conference, MOTOR , pages =. 2024 , doi =

    2024

  74. [74]

    Forty-first International Conference on Machine Learning,

    Ali Zeynali and Shahin Kamali and Mohammad Hajiesmaili , title =. Forty-first International Conference on Machine Learning,

  75. [75]

    Thirty-Fifth

    Ali Zeynali and Bo Sun and Mohammad Hassan Hajiesmaili and Adam Wierman , title =. Thirty-Fifth

  76. [76]

    1993 , isbn =

    Goldberg, Paul and Jerrum, Mark , title =. 1993 , isbn =. doi:10.1145/168304.168377 , booktitle =

    work page doi:10.1145/168304.168377 1993

  77. [77]

    2025 , organization =

    Citi Bike System Data , howpublished =. 2025 , organization =

    2025

  78. [78]

    Proceedings of the 35th Annual ACM Symposium on Theory of Computing (STOC 2003) , pages =

    Noga Alon and Baruch Awerbuch and Yossi Azar and Niv Buchbinder and Joseph Naor , title =. Proceedings of the 35th Annual ACM Symposium on Theory of Computing (STOC 2003) , pages =. 2003 , publisher =

    2003

  79. [79]

    Advances in Neural Information Processing Systems (NeurIPS 2022) , year =

    Anupam Gupta and Debmalya Panigrahi and Bernardo Subercaseaux and Kevin Sun , title =. Advances in Neural Information Processing Systems (NeurIPS 2022) , year =

    2022

  80. [80]

    Mathematical Optimization Theory and Operations Research , series =

    Yaroslav Kharchenko and Alexander Kononov , title =. Mathematical Optimization Theory and Operations Research , series =. 2024 , publisher =

    2024

Showing first 80 references.