Pith. sign in

REVIEW 3 major objections 4 minor 2 cited by

Improved Regret Analysis in Gaussian Process Bandits: Optimality for Noiseless Reward, RKHS norm, and Non-Stationary Variance

T0 review · 3 major / 4 minor · reviewed 2026-08-08 · deepseek-v4-flash

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

Pith's one-line read A sharper upper bound on the maximum posterior variance makes GP-bandit regret optimal in noiseless, RKHS-norm, and non-stationary-noise settings.

desk verdict A genuinely useful variance bound and three applications, but the main theorem leans on an unproved external count lemma; referee should demand a proof before acceptance. read the letter →

arxiv 2502.06363 v1 pith:I23JPW6P submitted 2025-02-10 cs.LG stat.ML

Shogo Iwazaki , Shion Takeno This is my paper · ORCID
classification cs.LGstat.ML MSC 68Q3262L0560G15
keywords GaussianprocessbanditsregretboundsmaximumvariancereductionphasedeliminationnoiselessrewardRKHSnormnon-stationarynoiseposterior
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

The paper's central aim is to sharpen Gaussian-process bandit regret bounds by proving a new uniform upper bound on the posterior standard deviation of a maximum-variance-reduction (MVR) query sequence. The bound improves how regret scales with the noise-variance parameter and with the RKHS norm of the reward, and it powers three sets of results: noiseless cumulative and simple regret that match or approach known lower bounds, RKHS-norm-optimal cumulative and simple regret, and the first regret analysis for GP bandits with time-varying observation noise. The main inequality says that after enough MVR queries the largest remaining uncertainty is at most of order $\sqrt{\tilde\lambda^2 \gamma_n / n}$, where $\gamma_n$ is the kernel's maximum information gain. A sympathetic reader would care because these are gaps that had been left open: the noiseless cumulative regret reaches the conjectured lower bound, and the non-stationary-noise guarantees are optimal up to logarithmic factors.

What carries the argument

The load-bearing object is Lemma 3.1, a uniform upper bound on the maximum posterior variance under MVR: for $n/2 \ge 3\gamma_n(\tilde\lambda_n^2 I_n)$, one has $\max_{x\in \tilde X} \sigma_{\lambda_n^2 I_n,n}(x; X_n) \le \frac{4}{n}\sqrt{\tilde\lambda_n^2 n \gamma_n(\tilde\lambda_n^2 I_n)}$, where $\gamma_n$ is the maximum information gain. The proof combines MVR's defining property—the current query point has the largest posterior variance, so the final maximum variance is no larger than the average over any subset of past queries—with an information-gain summation bound and Lemma 3.3, a quoted elliptical-potential count lemma that bounds the number of rounds whose scaled posterior standard deviation exceeds one. That count forces at least $n/2$ of the first $n$ queries to be informative, which converts the averaged variance into the displayed decay.

What would settle it

Run the MVR query sequence on a one-dimensional squared-exponential kernel with $\lambda_n^2 = \exp(-n^{1/(d+1)}\ln^{-\alpha} n)$, compute the set $T_> = \{t : \lambda^{-1}\sigma_{t-1}(x_t)>1\}$, and compare $|T_>|$ to the bound $3\gamma_n(\lambda^2 I)$. If $|T_>|$ exceeds that bound while $n/2 \ge 3\gamma_n(\lambda^2 I)$, then Lemma 3.3—and with it the regret theorems that rely on Lemma 3.1—fails.

Watch

Extended reading notes

Core claim

The core claim is that under the MVR selection rule, the maximum posterior standard deviation over the candidate set decays as $\sqrt{\tilde\lambda_n^2 \gamma_n / n}$ rather than at the slower rates previously known when the observation-variance parameter $\lambda^2$ tends to zero. This is Lemma 3.1, with Corollary 3.2 giving explicit rates for squared-exponential and Matérn kernels. From this inequality, phased elimination (PE) attains $O(\ln n)$ cumulative regret for squared-exponential kernels and the conjectured lower-bound rates for Matérn kernels in the noiseless setting, while MVR attains exponentially decaying simple regret for squared-exponential kernels and the order-optimal $n^{-\nu/d}$ rate for Matérn kernels. The same inequality yields simple- and cumulative-regret bounds with near-optimal dependence on the RKHS norm bound $B$ for noisy rewards. Applying it to a time-varying noise-variance model gives cumulative and simple regret bounds expressed through the cumulative variance proxy $V_n=\sum_{t}\sigma_t^2$, matching lower bounds derived from stationary and noiseless lower-bound arguments.

Load-bearing premise

A quoted counting lemma, not proved in the paper, must correctly bound how many rounds can have scaled posterior standard deviation above one; this count is what guarantees that at least half the MVR queries are informative, and every regret theorem collapses if the count is wrong.

Editorial extensions

If this is right

  • Noiseless cumulative regret: PE attains $O(\ln n)$ for squared-exponential kernels and the conjectured optimal rates for Matérn kernels, with deterministic guarantees rather than high-probability ones.
  • Noiseless simple regret: MVR attains $O(\exp(-\tfrac12 n^{1/(d+1)}\ln^{-\alpha} n))$ for squared-exponential kernels and $\tilde O(n^{-\nu/d})$ for Matérn kernels with $\nu>1/2$.
  • RKHS-norm optimality: setting the GP noise parameter to $\Theta(B^{-2})$ makes both simple and cumulative regret scale with the RKHS norm bound $B$ at the same polynomial rate as the lower bound, up to logarithmic factors, within the stated regimes of $B$-growth.
  • Non-stationary variance: VA-PE and VA-MVR regret bounds are governed by the cumulative variance proxy $V_n=\sum_{t}\sigma_t^2$ and match the derived lower bounds; when $V_n$ grows sublinearly, the algorithms beat the stationary-noise regret floor.

Reading between the lines

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

  • The same posterior-variance bound should transfer to Bayesian GP bandits with decreasing observation noise, yielding Bayesian regret improvements not explicitly derived in the paper.
  • If the conjectured MIG exponent of $\ln^{d/2}$ is established, the squared-exponential simple-regret exponent should improve from $-1/(d+1)$ toward $-2/d$, a strengthening the paper itself flags as plausible.
  • The non-stationary variance machinery could also apply to input-dependent heteroscedastic noise whenever the cumulative variance proxy grows sublinearly, suggesting a testable extension to drift-prone experimental settings that the paper only sketches.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper studies Gaussian process bandits and proposes a new upper bound on the maximum posterior variance under maximum variance reduction (MVR), stated as Lemma 3.1. The bound is then applied to three settings: noiseless rewards, regret bounds that are optimal in the RKHS norm upper bound B, and non-stationary observation noise variance. The authors prove regret upper bounds for PE- and MVR-style algorithms in these settings and provide lower bounds for the non-stationary variance setting. The main technical engine is Lemma 3.1, and most downstream results are derived as corollaries of it. The proof of the stationary case of Lemma 3.1 relies on an external elliptical-potential count lemma, Lemma 3.3, which is not proved in the manuscript.

Significance. If Lemma 3.1 and Lemma 3.3 are valid, the paper gives a unified treatment that improves on existing noise-variance dependence, recovers the conjectured noiseless Matérn lower-bound regime, obtains RKHS-norm-optimal simple and cumulative regret bounds under restricted growth of B, and provides the first regret analysis for non-stationary variance GP bandits. The non-stationary lower bounds and the explicit extension of the MIG-based argument to heteroscedastic noise are valuable contributions. However, the central count lemma is quoted from an unreviewed preprint, and the proof of the SE case of Theorem 5.1 has a condition mismatch; these issues must be resolved before the contributions can be fully assessed.

major comments (3)
  1. [Section 3, Lemma 3.3] Lemma 3.3 is load-bearing but is quoted from Flynn and Reeb [2024] and is not proved or even sketched in this manuscript. In the proof of Lemma 3.1 it is used at Eqs. (18)-(20) to obtain |T| >= n/2, and the refined form is also used in the proof of Theorem 5.3. Since Eq. (4) is the foundation for Corollary 3.2 and all theorems in Sections 4-6, the manuscript is not self-contained at its core. Please include a full proof of Lemma 3.3 or a precise verification that the MVR-induced sequences and the specific choices of lambda_tilde_n used in Sections 4-6 satisfy its hypotheses.
  2. [Theorem 5.1 / Appendix E.2] The SE case of Theorem 5.1 does not follow from the stated assumptions by the cited Corollary 3.2 statement 1. With lambda^2 = Theta(B^{-2}) and B = O(exp(n^{1/(d+1)} ln^{-alpha}(1+n))), one has lambda^2 = Omega(exp(-2 n^{1/(d+1)} ln^{-alpha}(1+n))), not lambda^2 = Omega(exp(-n^{1/(d+1)} ln^{-alpha}(1+n))). The proof's claim that the condition implies lambda^2 = O(exp(-(1/2) n^{1/(d+1)} ln^{-alpha}(1+n))) has the inequality reversed. The theorem may be recoverable by applying Lemma 3.1 directly, because ln(n/lambda^2) = O(n^{1/(d+1)} ln^{-alpha} n) makes gamma_n = o(n), but the proof as written is invalid and must be rewritten.
  3. [Lemma E.1 / Lemma F.1 and Theorems 5.3, 6.3] The confidence bounds in Lemma E.1 and Lemma F.1 are explicitly non-adaptive: they require the input sequence to be independent of the noise sequence. MVR satisfies this condition because its queries depend only on posterior variances, but PE and VA-PE eliminate candidates using noisy observations, so their subsequent batch designs are noise-dependent. The proofs of Theorems 5.3 and 6.3 invoke these lemmas without a conditioning argument. Please either prove a conditional version of the confidence bound for each batch or use a fully adaptive confidence bound; without this, the cumulative regret guarantees for PE and VA-PE are not established.
minor comments (4)
  1. [References] The references for Flynn and Reeb [2024] and Kim and Sanz-Alonso [2024] list the same arXiv identifier, arXiv:2401.17037; at least one of these identifiers is incorrect.
  2. [Table 2] The table mentions GP-UCB+ and EXPLOIT+ as algorithms from Kim and Sanz-Alonso, but these names are not introduced in the main text; a short explanation or pointer would help the reader.
  3. [Remark 4.3] Remark 4.3 states that the exact-argmax assumption can be relaxed with an extra sqrt(ln n) factor, but no formal theorem with the discretized variant is given; stating the precise assumptions and resulting bound would make the remark more useful.
  4. [Appendix C.1, proof of Lemma C.1] In Eq. (27) the notation tilde_X_{j-1} and the associated variance bound are clear, but the monotonicity step would benefit from an explicit sentence identifying that conditioning on fewer data points increases the posterior variance.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the regret bounds are derived from MVR, standard information-gain bounds, and an external count lemma; the unproved external lemma is a correctness risk, not a circular step.

full rationale

The derivation chain is not circular. Lemma 3.1 derives the maximum posterior variance bound from (i) the MVR selection rule, (ii) the standard sum-of-information-gain bound from Srinivas et al. (Eq. (17) uses I_{λ~^2I}(f(X),y) ≤ γ_n(λ~^2I)), and (iii) Lemma 3.3, the elliptical potential count lemma quoted from Flynn and Reeb [2024], used only to lower-bound |T|. None of these inputs is the target regret bound, and the target bound is not assumed in the definitions. Corollary 3.2 is an algebraic instantiation with known MIG bounds for SE and Matérn kernels. Theorems 4.1–6.4 are standard PE/MVR reductions that use Corollary 3.2 or Lemma 3.3, and the lower bounds they are compared with are quoted from Bull [2011], Scarlett et al. [2017], and Vakili [2022], none of which are authored by the present authors. The only self-citations (Takeno et al., 2023, 2024) appear in the related-work discussion of the Bayesian setting and are not load-bearing in any proof. No fitted parameter is renamed a prediction: λ^2 = Θ(B^{-2}) in Section 5 and λ_t = σ_t in Section 6 are design choices selected to match the noise model or to optimize the bound, not parameters fitted to the data whose regret is then claimed as a prediction. The one legitimate concern raised by the manuscript is that Lemma 3.3 is quoted from an external preprint and not proved or sketched in the paper; if that lemma or its hypotheses fail, Eqs. (4)–(5), Corollary 3.2, and all downstream regret bounds lose their foundation. This is an important correctness/verifiability risk, but it is not circular reasoning because the lemma is external, not an input defined in terms of the paper's outputs, and the paper's own equations do not reduce to their conclusions by construction.

Assumptions & free parameters 0 free parameters · 8 assumptions · 0 invented entities

No free parameters are fitted to data; the algorithm hyperparameters (λ²=Θ(B^{-2}), β, batch doubling) are design choices. No new physical or mathematical entities are posited. The load-bearing external inputs are standard modelling assumptions, known maximum information gain bounds, and one quoted count lemma.

assumptions (8)
  • domain assumption Assumption 2.1: f lies in the RKHS H_k of a known kernel with ||f||_H ≤ B and k(x,x) ≤ 1 for all x.
    This is the standard frequentist GP bandit assumption used in every confidence bound in the paper.
  • domain assumption Assumption 2.2: the noise sequence is mutually independent and sub-Gaussian with variance proxy σ_t².
    Needed for the concentration inequalities in Lemmas D.1, E.1, and F.1.
  • standard math Known maximum information gain bounds for SE and Matérn kernels: γ_n(λ²I)=O(ln^{d+1}(n/λ²)) and O((n/λ²)^{d/(2ν+d)} (ln(n/λ²))^{2ν/(2ν+d)}).
    Used to convert Lemma 3.1 into concrete kernel rates in Corollary 3.2 and Section 6; cited from prior work, not proved.
  • standard math Lemma 3.3 (elliptical potential count lemma, Lemma D.9 of Flynn and Reeb [2024]): |T_c| ≤ min{3γ(3γ_n(λ²I), λ²), 3γ_n(λ²I)}.
    Black-box lemma that gives |T|≥n/2 in the proof of Lemma 3.1 and appears again in Theorem 5.3; not proved in this paper.
  • standard math Lemma D.1: deterministic confidence bound for noiseless observations, |f(x)-μ(x)| ≤ Bσ(x) with λ=0.
    Attributed to Lyu et al. [2019] and Vakili et al. [2021]; underpins the noiseless regret theorems.
  • standard math Lemma E.1: non-adaptive confidence bound for noisy finite-X setting, with width B + (σ/λ) sqrt(2 ln(2|X|/δ)).
    Taken from Vakili et al. [2021]; used in all Section 5 simple and cumulative regret proofs.
  • domain assumption Exact argmax over continuous X in Algorithms 1 and 2 is assumed; Remark 4.3 notes discretization costs an extra sqrt(ln n) factor.
    The MVR selection identity max_x σ_n(x) ≤ average σ over queried points needs the exact argmax; continuous-domain exact optimization is computationally unrealistic but standard in this literature.
  • domain assumption Finite X is assumed in Theorems 5.1, 5.3, 6.3, and 6.4.
    Non-adaptive confidence bounds (Lemmas E.1 and F.1) use a union bound over |X|; finite X is stated in the theorems.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Improved Regret Analysis in Gaussian Process Bandits: Optimality for Noiseless Reward, RKHS norm, and Non-Stationary Variance." pith.science (2026). https://pith.science/paper/I23JPW6P

@misc{pith2026250206363,
  author       = {Pith},
  title        = {Pith review of: Improved Regret Analysis in Gaussian Process Bandits: Optimality for Noiseless Reward, RKHS norm, and Non-Stationary Variance},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/I23JPW6P}},
  note         = {Machine review of arXiv:2502.06363}
}
read the original abstract

We study the Gaussian process (GP) bandit problem, whose goal is to minimize regret under an unknown reward function lying in some reproducing kernel Hilbert space (RKHS). The maximum posterior variance analysis is vital in analyzing near-optimal GP bandit algorithms such as maximum variance reduction (MVR) and phased elimination (PE). Therefore, we first show the new upper bound of the maximum posterior variance, which improves the dependence of the noise variance parameters of the GP. By leveraging this result, we refine the MVR and PE to obtain (i) a nearly optimal regret upper bound in the noiseless setting and (ii) regret upper bounds that are optimal with respect to the RKHS norm of the reward function. Furthermore, as another application of our proposed bound, we analyze the GP bandit under the time-varying noise variance setting, which is the kernelized extension of the linear bandit with heteroscedastic noise. For this problem, we show that MVR and PE-based algorithms achieve noise variance-dependent regret upper bounds, which matches our regret lower bound.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Bayesian Optimization with Inexact Acquisition: Is Random Grid Search Sufficient?

    stat.ML 2025-06 conditional novelty 7.0 of 10

    Inexact acquisition maximization with bounded accumulated inaccuracy preserves sublinear regret, and random grid search with |X_t|=Theta(t) is a sufficient acquisition solver.

  2. No-Regret Gaussian Process Optimization of Time-Varying Functions

    stat.ML 2025-11 conditional novelty 6.0 of 10

    A windowed sparse GP-UCB with DPP-selected expert re-queries achieves sublinear dynamic regret using o(1) extra queries per round on average, and a Fano lower bound shows Ω(T^{α/(α+1)}) queries are needed in fast-drif...

Reference graph

Works this paper leans on

34 extracted references · 32 canonical work pages · cited by 2 Pith papers

  1. [1]

    Online learning for linearly parametrized control problems

    Yasin Abbasi-Yadkori. Online learning for linearly parametrized control problems. PhD thesis, University of Alberta, 2013

    2013

  2. [2]

    Schoellig

    Felix Berkenkamp, Andreas Krause, and Angela P. Schoellig. B ayesian optimization with safety constraints: safe and automatic parameter tuning in robotics. Mach. Learn., 112 0 (10): 0 3713^^e2^^80^^933747, 2021

    work page 2021

  3. [3]

    Convergence rates of efficient global optimization algorithms

    Adam D Bull. Convergence rates of efficient global optimization algorithms. Journal of Machine Learning Research, 2011

    work page 2011

  4. [4]

    On lower bounds for standard and robust G aussian process bandit optimization

    Xu Cai and Jonathan Scarlett. On lower bounds for standard and robust G aussian process bandit optimization. In Proceedings of the 38th International Conference on Machine Learning, volume 139 of Proceedings of Machine Learning Research, pages 1216--1226. PMLR, 2021

    work page 2021

  5. [5]

    High-dimensional experimental design and kernel bandits

    Romain Camilleri, Kevin Jamieson, and Julian Katz-Samuels. High-dimensional experimental design and kernel bandits. In Proceedings of the 38th International Conference on Machine Learning, volume 139 of Proceedings of Machine Learning Research, pages 1227--1237. PMLR, 2021

    work page 2021

  6. [6]

    On kernelized multi-armed bandits

    Sayak Ray Chowdhury and Aditya Gopalan. On kernelized multi-armed bandits. In Proc. International Conference on Machine Learning (ICML), 2017

    work page 2017

  7. [7]

    Cover and Joy A

    Thomas M. Cover and Joy A. Thomas. Elements of Information Theory (Wiley Series in Telecommunications and Signal Processing). Wiley-Interscience, 2006

    work page 2006

  8. [8]

    Smola, and Masrour Zoghi

    Nando De Freitas, Alex J. Smola, and Masrour Zoghi. Exponential regret bounds for G aussian process bandits with deterministic observations. In Proceedings of the 29th International Conference on International Conference on Machine Learning, page 955^^e2^^80^^93962. Omnipress, 2012

    work page 2012

Show all 34 references
  1. [10]

    Rapid, accurate, and precise concentration measurements of a methanol--water mixture using R aman spectroscopy

    Daniel D Hickstein, Russell Goldfarbmuren, Jack Darrah, Luke Erickson, and Laura A Johnson. Rapid, accurate, and precise concentration measurements of a methanol--water mixture using R aman spectroscopy. OSA Continuum, 1 0 (3): 0 1097--1110, 2018

    2018

  2. [11]

    Bandit optimisation of functions in the M at^^c3^^a9rn kernel RKHS

    David Janz, David Burt, and Javier Gonzalez. Bandit optimisation of functions in the M at^^c3^^a9rn kernel RKHS . In Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics, volume 108 of Proceedings of Machine Learning Research, page...

    2020

  3. [12]

    Enhancing G aussian process surrogates for optimization and posterior approximation via random exploration

    Hwanwoo Kim and Daniel Sanz-Alonso. Enhancing G aussian process surrogates for optimization and posterior approximation via random exploration. arXiv preprint arXiv:2401.17037, 2024

    2024 arXiv

  4. [13]

    Improved regret analysis for variance-adaptive linear bandits and horizon-free linear mixture MDP s

    Yeoneung Kim, Insoon Yang, and Kwang-Sung Jun. Improved regret analysis for variance-adaptive linear bandits and horizon-free linear mixture MDP s. In Advances in Neural Information Processing Systems, volume 35, pages 1060--1072. Curran Associates, Inc., 2022

    2022

  5. [14]

    Information directed sampling and bandits with heteroscedastic noise

    Johannes Kirschner and Andreas Krause. Information directed sampling and bandits with heteroscedastic noise. In Proc. Conference on Learning Theory (COLT), 2018

    2018

  6. [15]

    Multi-scale zero-order optimization of smooth functions in an RKHS

    Madison Lee, Shubhanshu Shekhar, and Tara Javidi. Multi-scale zero-order optimization of smooth functions in an RKHS . In 2022 IEEE International Symposium on Information Theory (ISIT). IEEE, 2022

    2022

  7. [16]

    B ayesian optimization with adaptive surrogate models for automated experimental design

    Bowen Lei, Tanner Quinn Kirk, Anirban Bhattacharya, Debdeep Pati, Xiaoning Qian, Raymundo Arroyave, and Bani K Mallick. B ayesian optimization with adaptive surrogate models for automated experimental design. npj Computational Materials, 7 0 (1): 0 194, 2021

    2021

  8. [17]

    G aussian process bandit optimization with few batches

    Zihan Li and Jonathan Scarlett. G aussian process bandit optimization with few batches. In Proc. International Conference on Artificial Intelligence and Statistics (AISTATS), 2022

    2022

  9. [18]

    Efficient batch black-box optimization with deterministic regret bounds

    Yueming Lyu, Yuan Yuan, and Ivor W Tsang. Efficient batch black-box optimization with deterministic regret bounds. arXiv preprint arXiv:1905.10041, 2019

    1905 arXiv

  10. [19]

    Risk-averse heteroscedastic B ayesian optimization

    Anastasia Makarova, Ilnura Usmanova, Ilija Bogunovic, and Andreas Krause. Risk-averse heteroscedastic B ayesian optimization. In Proc. Neural Information Processing Systems (NeurIPS), 2021

    2021

  11. [20]

    Learning to optimize via posterior sampling

    Daniel Russo and Benjamin Van Roy. Learning to optimize via posterior sampling. Mathematics of Operations Research, 39 0 (4): 0 1221--1243, 2014

    2014

  12. [21]

    A domain-shrinking based B ayesian optimization algorithm with order-optimal regret performance

    Sudeep Salgia, Sattar Vakili, and Qing Zhao. A domain-shrinking based B ayesian optimization algorithm with order-optimal regret performance. In Advances in Neural Information Processing Systems, volume 34, pages 28836--28847. Curran Associates, Inc., 2021

    2021

  13. [22]

    Random exploration in B ayesian optimization: O rder-optimal regret and computational efficiency

    Sudeep Salgia, Sattar Vakili, and Qing Zhao. Random exploration in B ayesian optimization: O rder-optimal regret and computational efficiency. In Proc. International Conference on Machine Learning (ICML), 2024

    2024

  14. [23]

    Tight regret bounds for B ayesian optimization in one dimension

    Jonathan Scarlett. Tight regret bounds for B ayesian optimization in one dimension. In Proceedings of the 35th International Conference on Machine Learning, volume 80 of Proceedings of Machine Learning Research, pages 4500--4508. PMLR, 2018

    2018

  15. [24]

    Lower bounds on regret for noisy G aussian process bandit optimization

    Jonathan Scarlett, Ilija Bogunovic, and Volkan Cevher. Lower bounds on regret for noisy G aussian process bandit optimization. In Proc. Conference on Learning Theory (COLT), 2017

    2017

  16. [25]

    Practical B ayesian optimization of machine learning algorithms

    Jasper Snoek, Hugo Larochelle, and Ryan P Adams. Practical B ayesian optimization of machine learning algorithms. In Proc. Neural Information Processing Systems (NeurIPS), 2012

    2012

  17. [26]

    Gaussian process optimization in the bandit setting: No regret and experimental design

    Niranjan Srinivas, Andreas Krause, Sham Kakade, and Matthias Seeger. Gaussian process optimization in the bandit setting: No regret and experimental design. In Proc. International Conference on Machine Learning (ICML), 2010

    2010

  18. [27]

    Randomized G aussian process upper confidence bound with tighter B ayesian regret bounds

    Shion Takeno, Yu Inatsu, and Masayuki Karasuyama. Randomized G aussian process upper confidence bound with tighter B ayesian regret bounds. In Proceedings of the 40th International Conference on Machine Learning, volume 202 of Proceedings of Machine Learning Research, pages 33...

    2023

  19. [28]

    Posterior sampling-based B ayesian optimization with tighter B ayesian regret bounds

    Shion Takeno, Yu Inatsu, Masayuki Karasuyama, and Ichiro Takeuchi. Posterior sampling-based B ayesian optimization with tighter B ayesian regret bounds. In Proceedings of the 41st International Conference on Machine Learning, volume 235 of Proceedings of Machine Learning Resea...

    2024

  20. [29]

    Open problem: R egret bounds for noise-free kernel-based bandits

    Sattar Vakili. Open problem: R egret bounds for noise-free kernel-based bandits. In Proc. Conference on Learning Theory (COLT), 2022

    2022

  21. [30]

    Optimal order simple regret for G aussian process bandits

    Sattar Vakili, Nacime Bouziani, Sepehr Jalali, Alberto Bernacchia, and Da shan Shiu. Optimal order simple regret for G aussian process bandits. In Proc. Neural Information Processing Systems (NeurIPS), 2021

    2021

  22. [31]

    Finite-time analysis of kernelised contextual bandits

    Michal Valko, Nathan Korda, R\' e mi Munos, Ilias Flaounas, and Nello Cristianini. Finite-time analysis of kernelised contextual bandits. In Proceedings of the Twenty-Ninth Conference on Uncertainty in Artificial Intelligence, UAI'13, page 654^^e2^^80^^93663. AUAI Press, 2013

    2013

  23. [32]

    Improved variance-aware confidence sets for linear bandits and linear mixture MDP

    Zihan Zhang, Jiaqi Yang, Xiangyang Ji, and Simon S Du. Improved variance-aware confidence sets for linear bandits and linear mixture MDP . In Advances in Neural Information Processing Systems, volume 34, pages 4342--4355. Curran Associates, Inc., 2021

    2021

  24. [33]

    Variance-dependent regret bounds for linear bandits and reinforcement learning: A daptivity and computational efficiency

    Heyang Zhao, Jiafan He, Dongruo Zhou, Tong Zhang, and Quanquan Gu. Variance-dependent regret bounds for linear bandits and reinforcement learning: A daptivity and computational efficiency. In Proc. Conference on Learning Theory (COLT). PMLR, 2023

    2023

  25. [34]

    Computationally efficient horizon-free reinforcement learning for linear mixture MDP s

    Dongruo Zhou and Quanquan Gu. Computationally efficient horizon-free reinforcement learning for linear mixture MDP s. In Advances in Neural Information Processing Systems, volume 35, pages 36337--36349. Curran Associates, Inc., 2022

    2022

  26. [35]

    Nearly minimax optimal reinforcement learning for linear mixture M arkov decision processes

    Dongruo Zhou, Quanquan Gu, and Csaba Szepesvari. Nearly minimax optimal reinforcement learning for linear mixture M arkov decision processes. In Proceedings of Thirty Fourth Conference on Learning Theory, volume 134 of Proceedings of Machine Learning Research, pages 4532--4576...

    2021

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.