Pith. sign in

REVIEW 3 major objections 5 minor 2 cited by

Online Experimental Design With Estimation-Regret Trade-off Under Network Interference

T0 review · 3 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A unified exposure-mapping framework for online experiments proves that regret and ATE estimation under network interference obey the minimax trade-off $\sqrt{R e} = \Omega(\sqrt{|U_E|})$, and the two-stage UCB-TSN algorithm achieves the…

desk verdict The MAB-N framework and matching scalar trade-off are solid and worth citing, but the Pareto-optimality theorem is not proved as written. read the letter →

arxiv 2412.03727 v3 pith:IJEPJ5SX submitted 2024-12-04 cs.LG math.OCmath.STstat.TH

Zhiheng Zhang , Zichen Wang This is my paper · ORCID
classification cs.LGmath.OCmath.STstat.TH
keywords multi-armedbanditsnetworkinterferenceexposuremappingaveragetreatmenteffectregret-estimationtrade-offParetooptimalityonlineexperimentaldesigntwo-stageUCB
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

Online experiments over networked populations face two goals that pull in opposite directions: estimating average treatment effects requires revisiting suboptimal assignments, while minimizing regret requires abandoning them. The paper claims that under network interference this tension is governed by the size of a compressed action space: for any policy, the product of cumulative regret $R$ and worst-case ATE estimation error $e$ is at least $\Omega(\sqrt{|U_E|})$, where $U_E$ is the exposure arm space obtained from the exposure mapping. It then gives a two-stage algorithm, UCB-TSN, whose product is $\widetilde{O}(\sqrt{|U_E|})$, so the trade-off is Pareto-optimal. The result holds for general network topologies under a mild condition, and its scope includes the no-interference case, where it improves the previously known trade-off.

What carries the argument

The load-bearing object is the exposure mapping $S(i,A,H)$, a pre-specified function that summarizes the full treatment assignment $A$ into a low-cardinality exposure arm for each unit; the whole network state is then represented by the exposure super arm $S_t$, living in the set $U_E$ of cluster-wise switchback exposures that are also realizable by some original assignment. This reduces the exponentially large super-arm space $K^N$ to $|U_E| \le |\mathcal{U}_s|^C$ and makes the online problem tractable. The proof machinery is the divergence bound $KL(P_{\nu_1},P_{\nu_2}) \le qN\alpha^2 E_{\nu_1}[N_T^S]$, which ties the distance between two hard instances to the expected number of pulls of a suboptimal exposure arm, converting estimation difficulty directly into regret. On the algorithmic side, UCB-TSN's first phase (round-robin over $U_E$) supplies the uniform sampling that makes ATE estimation consistent, and its second phase (UCB on $U_E$) controls regret; the choice $T_1=\sqrt{|U_E|T}$ balances the two bounds.

What would settle it

The proof of Theorem 4 hinges on the divergence bound $KL(P_{\nu_1},P_{\nu_2}) \le qN\alpha^2 E_{\nu_1}[N_T^S]$. A reader can test the hinge on a two-unit network with $|U_E|=2$: if the KL divergence between the two constructed instances grows faster than $\alpha^2$ times the expected number of pulls of the suboptimal exposure arm, the lower bound collapses. A complementary stress test is to run UCB-TSN with an exposure mapping that leaves out a real neighbor effect and check whether regret measured against the true optimal assignment stays $\widetilde{O}(\sqrt{|U_E|T})$.

Watch

Extended reading notes

Core claim

The central claim, established as Theorem 4 and Corollary 8, is that regret minimization and ATE estimation cannot both be pushed below a common barrier in online experiments with network interference. For any online policy and any estimator, the paper constructs two statistically close reward instances whose only meaningful difference is the ATE of a chosen exposure arm; the estimation error of any estimator and the number of pulls of that arm are coupled through a KL-divergence calculation, yielding $\inf_{\hat\Delta}\max_\nu \sqrt{R_\nu(T,\pi)e_\nu(T,\hat\Delta)} = \Omega(\sqrt{|U_E|})$ up to constants in $K$ and $T$. A pair is Pareto-optimal whenever this product is $\widetilde{O}(\sqrt{|U_E|})$, and UCB-TSN attains exactly that by exploring all exposure arms uniformly for $T_1 = \sqrt{|U_E|T}$ rounds and then running UCB on the exposure arm set. This both subsumes and strengthens the earlier estimation-regret trade-off obtained without interference.

Load-bearing premise

The exposure mapping is assumed to capture all ways treatment reaches outcomes: the potential outcome of every unit depends on the full assignment only through the exposure super arm $S_t$, and the regret target $S^*$ is defined in that reduced space. If a real interference channel is left out of $S$, the guarantees attach to a misspecified best arm.

Editorial extensions

If this is right

  • For any network topology and any exposure mapping satisfying $2\le |U_E|\le T$, no algorithm can reduce the product of regret and ATE estimation error below $\Omega(\sqrt{|U_E|})$.
  • UCB-TSN reaches the frontier: with $T_1=\sqrt{|U_E|T}$ it guarantees $R=\widetilde{O}(\sqrt{|U_E|T})$ and $e=\widetilde{O}(\sqrt{|U_E|/T_1})$, so $\sqrt{Re}=\widetilde{O}(\sqrt{|U_E|})$.
  • Practitioners can dial between estimation and regret by choosing $T_1$ along the Pareto frontier, since $T_1=0$ recovers standard UCB-like regret while $T_1=T$ recovers uniform estimation design.
  • In the single-unit, no-interference special case, the regret bound becomes $\widetilde{O}(K+T_1)$ and the estimation bound $\widetilde{O}(\sqrt{K/T_1})$, which is stronger than the earlier baseline trade-off.
  • Replacing UCB with an EXP3 variant extends the same Pareto-optimal trade-off to adversarial rewards.

Reading between the lines

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

  • Beyond the paper: the lower bound identifies $|U_E|$, not the number of units or original arms, as the complexity measure that governs the regret-estimation product, so any refinement of the exposure mapping or clustering that enlarges $U_E$ carries a direct statistical cost.
  • Beyond the paper: if the exposure mapping misses a real interference channel, the guarantees concern the best exposure arm rather than the true best assignment; a concrete stress test would run UCB-TSN on a network where an unobserved neighbor's treatment drives outcomes and check whether regret relative to the true optimum stays sublinear.
  • Beyond the paper: the paper's Appendix L replaces exact compatibility with Manhattan-distance proximity, but no analogous Pareto-optimality bound is proved there; proving one for the approximate case is a natural next step.
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 / 5 minor

Summary. The paper proposes MAB-N, a stochastic multi-armed bandit framework under network interference, where the action space is reduced via a statistical exposure mapping and clustering. The central claims are: (i) a minimax lower bound, inf_{Δ̂} max_ν sqrt(R_ν e_ν) = Ω(sqrt(|U_E|)) for any policy π (Theorem 4); (ii) a sufficiency criterion asserting that achieving max_ν sqrt(R_ν e_ν) = O~(sqrt(|U_E|)) implies Pareto-optimality (Theorem 5); and (iii) an algorithm UCB-TSN, whose two-stage design matches the lower bound (Corollary 8). An extension to an adversarial setting via EXP3-TSN is also given. The paper interprets these results as the first Pareto-optimal estimation-regret trade-off under general network interference.

Significance. If the claims were fully established, the paper would be a solid contribution: it unifies several prior bandit-under-interference models as special cases of a single exposure-mapping framework, and it extends the estimation-regret trade-off of Simchi-Levi and Wang (2024) from a single unit to general networks. The two-instance minimax construction and the UCB-type upper bound are standard in spirit and appear mostly coherent. However, the headline Pareto-optimality claim rests on Theorem 5, whose proof is invalid, and the proof of Theorem 4 also contains a gap in the translation from E[max] to max E. These are load-bearing issues, not cosmetic ones. The paper also includes reproducible-style algorithm pseudocode and a conservative choice of sub-Gaussian proxy (σ²=2) that keeps the concentration bounds safe despite an erroneous derivation in Appendix E.

major comments (3)
  1. [§4, Theorem 5 and Appendix H] The sufficiency proof is a non sequitur. From Pareto-domination of {π,Δ̂} by {π',Δ̂'}, the proof asserts that Theorem 4 gives an instance ν' with sqrt(R_{ν'}(π') e_{ν'}(Δ̂')) = Ω(sqrt(|U_E|)), and then that 'there further exists another instance ν''' with sqrt(R_{ν''}(π)e_{ν''}(Δ̂)) strictly larger than sqrt(|U_E|). The second step does not follow: Pareto-domination compares front points F(π,Δ̂), and the hard instance ν' for π' need not lie on the front of π', nor does it force any instance on the front of π to have large product. The scalarized worst-case product can be attained at a dominated, off-front instance, so the bound max_ν sqrt(R_ν e_ν) = O~(sqrt(|U_E|)) does not preclude front domination by another algorithm. Consequently, the claim that UCB-TSN is Pareto-optimal (Corollary 8, abstract) is not established by the given argument.
  2. [§4, Theorem 4 proof, Appendix G, Eq. (28)] The proof lower-bounds E_ν[max_{i,j}|Δ̂^{(i,j)}_t - Δ^{(i,j)}_ν|] and multiplies by sqrt(R_ν(T,π)), but the estimation error e_ν is defined in Eq. (5) as max_{i,j} E_ν[|Δ̂^{(i,j)}_t - Δ^{(i,j)}_ν|]. Since max_i E ≤ E max_i, the displayed lower bound on E[max] does not imply the claimed lower bound on e_ν. The argument can be repaired by applying the hypothesis-testing lower bound to the specific pair (S,S') selected in the construction and using E|Δ̂^{(S,S')} − Δ^{(S,S')}| ≥ (α/2)P(|Δ̂^{(S,S')} − Δ^{(S,S')}| ≥ α/2); as written, the proof does not establish Theorem 4.
  3. [Appendix E, Eqs. (8)–(10)] The derivation of the 'natural' sub-Gaussian proxy uses cosh(m/2) ≤ exp(m²/8) for Z = (1/N)∑_i (Y_i(A) − Ỹ_i(S_t)). This Z is zero-mean and bounded in [−1,1], so its MGF can be as large as cosh(m), not cosh(m/2); the claimed proxy 1/N + 1/4 is not justified. A valid bound for the sampling-noise term alone gives an additional factor exp(m²/2), so the combined proxy is at most 1/N + 1/2. Since the paper subsequently uses the conservative σ² = 2, the concentration bounds in Appendix I remain valid, but the derivation in Appendix E and the remarks about the order of N should be corrected.
minor comments (5)
  1. [Definition 2, condition (ii)] The direction of the domination relation appears reversed: as written, the condition forbids an instance ν for which (R,e) is componentwise ≤ (sqrt(R_ν),e_ν), which would define maximal rather than minimal front points; the intended lower-left envelope requires the opposite inequality.
  2. [Theorem 5 statement] The claimed set equality 'Pareto Frontier is represented as P = {(R_ν,e_ν) : sqrt(R_ν e_ν) = O~(sqrt(|U_E|))}' uses asymptotic notation inside a set definition; the intended asymptotic characterization should be stated explicitly.
  3. [Algorithm 1] In the pseudocode, the second-phase loop does not explicitly show the update of N_t^S and R̂_t(S); the 'Parameter' lines should be placed inside the for-loop to match the stated update rules.
  4. [Appendix G, Eqs. (27)–(28)] The notation Δ̂_t in the probability bound should be indexed by the specific pair (S,S') used in the construction, to avoid the E[max] versus max E ambiguity in the conversion to a bound on e_ν.
  5. [Appendix G, Eq. (30)] When choosing α so that qα²N R_{ν1}(T,π)/(|U_E||Δ1|) = 1/2, the proof should state explicitly that Condition 1 (T ≥ |U_E|) is used and should verify that α ≤ 1 remains feasible given the lower bound on |Δ1|.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the regret/ATE trade-off bounds are independently derived; the sufficiency proof of Theorem 5 has a non-circular proof gap.

full rationale

The central scalarized trade-off is not circular. Theorem 4 is proved in Appendix G by a standard two-instance minimax construction: instances ν1 and ν2 are engineered so that the ATE gap differs by exactly −α (Eq. 19), the KL divergence is bounded by qNα²E[N_S] (Eqs. 24–26), and the final bound follows from choosing α to balance the terms. No step of that proof assumes the target regret/estimation trade-off, and no fitted parameter is later renamed as a prediction. The upper-bound side is also self-contained: Theorem 6 follows from round-robin exploration plus Hoeffding/union-bound concentration, and Theorem 7 follows from the UCB event analysis and the choice T1 = sqrt(|UE|T). Corollary 8 multiplies the two bounds; it is algebra, not a fit. The lower and upper bounds match in order, so the scalarized statement sqrt(R·e) = Θ~(sqrt(|UE|)) is independently established. The remaining 'Pareto-optimality' claim relies on Theorem 5, whose proof in Appendix H contains a non-sequitur: from Pareto-domination and Theorem 4 it asserts 'there further exists another instance ν″' with sqrt(|UE|) ≺ sqrt(R_{ν″}(π)e_{ν″}(Δ)); this does not follow because the hard instance supplied by Theorem 4 for the dominating pair need not lie on that pair's front, and domination of fronts does not imply a single instance on which the dominated pair exceeds the worst-case product. This is a correctness gap in a derived implication, not a circular reduction: Definition 2 does not define Pareto-optimality as the scalar product bound, and the attempted proof does not assume its conclusion. Since the load-bearing numerical trade-off is derived independently, the circularity score is 1; this should not be read as endorsing Theorem 5's validity.

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

The framework rests on standard MAB and exposure-mapping assumptions; the only ad hoc element is the loose sub-Gaussian proxy, which is conservative and does not affect the order of the final bounds. No free parameters or invented entities are introduced.

assumptions (4)
  • domain assumption Potential outcomes Yi(A) lie in [0,1] and noise eta_i,t is zero-mean 1-sub-Gaussian and independent across units and rounds
    Used throughout to apply Hoeffding/Bernstein concentration and to define the bandit instance class E0 (Section 3).
  • domain assumption Exposure mapping S and clusters C are pre-specified, and the induced exposure super arm U_E satisfies Condition 1: 2 <= |U_E| <= T
    Condition 1 is required in Theorems 4-9; it restricts the regime where the trade-off results hold.
  • domain assumption The potential outcome depends on the assignment only through the exposure super arm, so Ytilde_i(S) := Sum_A Yi(A) P(A|S) is the true reward for exposure arm S
    This is the exposure-mapping identification assumption (Eq (2)-(3)), needed for the regret in Eq (4) to be a meaningful objective; misspecification is only informally discussed in Appendix L.
  • ad hoc to paper The average exposure reward is sigma^2=2 sub-Gaussian
    Appendix E derives a proxy but with a loose step (cosh(m/2) instead of cosh(m)); the final sigma^2=2 is conservative enough if rewards are bounded, but the derivation in the paper is not fully correct.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Online Experimental Design With Estimation-Regret Trade-off Under Network Interference." pith.science (2026). https://pith.science/paper/IJEPJ5SX

@misc{pith2026241203727,
  author       = {Pith},
  title        = {Pith review of: Online Experimental Design With Estimation-Regret Trade-off Under Network Interference},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IJEPJ5SX}},
  note         = {Machine review of arXiv:2412.03727}
}
read the original abstract

Network interference has attracted significant attention in the field of causal inference, encapsulating various sociological behaviors where the treatment assigned to one individual within a network may affect the outcomes of others, such as their neighbors. A key challenge in this setting is that standard causal inference methods often assume independent treatment effects among individuals, which may not hold in networked environments. To estimate interference-aware causal effects, a traditional approach is to inherit the independent settings, where practitioners randomly assign experimental participants into different groups and compare their outcomes. While effective in offline settings, this strategy becomes problematic in sequential experiments, where suboptimal decision persists, leading to substantial regret. To address this issue, we introduce a unified interference-aware framework for online experimental design. Compared to existing studies, we extend the definition of arm space by utilizing the statistical concept of exposure mapping, which allows for a more flexible and context-aware representation of treatment effects in networked settings. Crucially, we establish a Pareto-optimal trade-off between estimation accuracy and regret under the network concerning both time period and arm space, which remains superior to baseline models even without network interference. Furthermore, we propose an algorithmic implementation and discuss its generalization across different learning settings and network topology.

Figures

Figures reproduced from arXiv: 2412.03727 by the authors.

Figure 1
Figure 1. Pareto-optimality. (a) We use three blue fronts (first quadrant) to show three [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 2
Figure 2. Network structure. Setup. We consider a network consisting of 101 units. Specifically, there is a central cluster C1 = {1} that contains a single unit, which is connected to every unit in the five peripheral clusters C2, . . . , C6 (namely, C2 = {2, . . . , 21}, C3 = {22, . . . , 41}, C4 = {42, . . . , 61}, C5 = {62, . . . , 81}, and C6 = {82, . . . , 101}, with each outer cluster containing 20 units, as shown in [… view at source ↗
Figure 3
Figure 3. Experimental results. Gao and Ding, 2023), we define the exposure mapping as S(i, A, H) = 1 nP j P hijaj j hij ∈ [PITH_FULL_IMAGE:figures/full_fig_p025_3.png] view at source ↗

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. Balancing Interference and Correlation in Spatial Experimental Designs: A Causal Graph Cut Approach

    cs.LG 2025-05 conditional novelty 7.0 of 10

    A surrogate for the ATE estimator's MSE is optimized with spectral graph cuts to produce cluster-randomized designs that adapt to the spatial covariance and accommodate moderate-to-large interference.

  2. Learning Peer Influence Probabilities with Linear Contextual Bandits

    cs.LG 2025-10 conditional novelty 6.0 of 10

    In a linear contextual bandit setting with k network interventions per round, cumulative regret and influence-probability estimation error obey a rate trade-off, and a new algorithm, InfluenceCB, can attain any point ...

Reference graph

Works this paper leans on

65 extracted references · 42 canonical work pages · cited by 2 Pith papers

  1. [1]

    Multi-armed bandits with network interference

    Abhineet Agarwal, Anish Agarwal, Lorenzo Masoero, and Justin Whitehouse. Multi-armed bandits with network interference. arXiv preprint arXiv:2405.18621, 2024

    arXiv 2024

  2. [2]

    Thompson sampling for contextual bandits with linear payoffs

    Shipra Agrawal and Navin Goyal. Thompson sampling for contextual bandits with linear payoffs. In International Conference on Machine Learning, 2012

    work page 2012

  3. [3]

    Asymptotically efficient allocation rules for the multiarmed bandit problem with multiple plays-part i: Iid rewards

    Venkatachalam Anantharam, Pravin Varaiya, and Jean Walrand. Asymptotically efficient allocation rules for the multiarmed bandit problem with multiple plays-part i: Iid rewards. IEEE Transactions on Automatic Control, 32 0 (11): 0 968--976, 1987

    work page 1987

  4. [4]

    Estimating average causal effects under general interference, with application to a social network experiment

    Peter M Aronow and Cyrus Samii. Estimating average causal effects under general interference, with application to a social network experiment. The Annals of Applied Statistics, 2017

    work page 2017

  5. [5]

    Assessing the causal effects of financial aids to firms in tuscany allowing for interference

    Bruno Arpino and Alessandra Mattei. Assessing the causal effects of financial aids to firms in tuscany allowing for interference. The Annals of Applied Statistics, 2016

    work page 2016

  6. [6]

    Finite-time analysis of the multiarmed bandit problem

    Peter Auer, Nicol \`o Cesa-Bianchi, and Paul Fischer. Finite-time analysis of the multiarmed bandit problem. Machine Learning, 47: 0 235--256, 2002

    work page 2002

  7. [7]

    Social connections and incentives in the workplace: Evidence from personnel data

    Oriana Bandiera, Iwan Barankay, and Imran Rasul. Social connections and incentives in the workplace: Evidence from personnel data. Econometrica, 77 0 (4): 0 1047--1094, 2009

    work page 2009

  8. [8]

    Multi-player bandits revisited

    Lilian Besson and Emilie Kaufmann. Multi-player bandits revisited. In Algorithmic Learning Theory, pages 56--92. PMLR, 2018

    work page 2018

Show all 65 references
  1. [9]

    A 61-million-person experiment in social influence and political mobilization

    Robert M Bond, Christopher J Fariss, Jason J Jones, Adam DI Kramer, Cameron Marlow, Jaime E Settle, and James H Fowler. A 61-million-person experiment in social influence and political mobilization. Nature, 489 0 (7415): 0 295--298, 2012

    2012

  2. [10]

    Social networks and the decision to insure

    Jing Cai, Alain De Janvry, and Elisabeth Sadoulet. Social networks and the decision to insure. American Economic Journal: Applied Economics, 7: 0 81--108, 2015

    2015

  3. [11]

    Combinatorial bandits

    Nicolo Cesa-Bianchi and G \'a bor Lugosi. Combinatorial bandits. Journal of Computer and System Sciences, 78 0 (5): 0 1404--1422, 2012

    2012

  4. [12]

    Combinatorial pure exploration of multi-armed bandits

    Shouyuan Chen, Tian Lin, Irwin King, Michael R Lyu, and Wei Chen. Combinatorial pure exploration of multi-armed bandits. Advances in Neural Information Processing Systems, 27, 2014

    2014

  5. [13]

    Combinatorial multi-armed bandit: General framework and applications

    Wei Chen, Yajun Wang, and Yang Yuan. Combinatorial multi-armed bandit: General framework and applications. In International conference on machine learning, pages 151--159. PMLR, 2013

    2013

  6. [14]

    Reward maximization for pure exploration: Minimax optimal good arm identification for nonparametric multi-armed bandits

    Brian Cho, Dominik Meier, Kyra Gan, and Nathan Kallus. Reward maximization for pure exploration: Minimax optimal good arm identification for nonparametric multi-armed bandits. arXiv preprint arXiv:2410.15564, 2024

    2024 arXiv

  7. [15]

    The covid-19 pandemic

    Marco Ciotti, Massimo Ciccozzi, Alessandro Terrinoni, Wen-Can Jiang, Cheng-Bin Wang, and Sergio Bernardini. The covid-19 pandemic. Critical reviews in clinical laboratory sciences, 57 0 (6): 0 365--388, 2020

    2020

  8. [16]

    Combinatorial bandits revisited

    Richard Combes, Mohammad Sadegh Talebi Mazraeh Shahi, Alexandre Proutiere, et al. Combinatorial bandits revisited. Advances in Neural Information Processing Systems, 28, 2015

    2015

  9. [17]

    Online debiasing for adaptively collected high-dimensional data with applications to time series analysis

    Yash Deshpande, Adel Javanmard, and Mohammad Mehrabi. Online debiasing for adaptively collected high-dimensional data with applications to time series analysis. Journal of the American Statistical Association, 118 0 (542): 0 1126--1139, 2023

    2023

  10. [18]

    Estimation considerations in contextual bandits

    Maria Dimakopoulou, Zhengyuan Zhou, Susan Athey, and Guido Imbens. Estimation considerations in contextual bandits. arXiv preprint arXiv:1711.07077, 2017

    2017 arXiv

  11. [19]

    Balanced linear contextual bandits

    Maria Dimakopoulou, Zhengyuan Zhou, Susan Athey, and Guido Imbens. Balanced linear contextual bandits. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 33, pages 3445--3453, 2019

    2019

  12. [20]

    Online multi-armed bandits with adaptive inference

    Maria Dimakopoulou, Zhimei Ren, and Zhengyuan Zhou. Online multi-armed bandits with adaptive inference. Advances in Neural Information Processing Systems, 34: 0 1939--1951, 2021

    1939

  13. [21]

    Regret minimization and statistical inference in online decision making with high-dimensional covariates

    Congyuan Duan, Wanteng Ma, Jiashuo Jiang, and Dong Xia. Regret minimization and statistical inference in online decision making with high-dimensional covariates. arXiv preprint arXiv:2411.06329, 2024

    2024 arXiv

  14. [22]

    Overlap in observational studies with high-dimensional covariates

    Alexander D’Amour, Peng Ding, Avi Feller, Lihua Lei, and Jasjeet Sekhon. Overlap in observational studies with high-dimensional covariates. Journal of Econometrics, 221 0 (2): 0 644--654, 2021

    2021

  15. [23]

    Trading off rewards and errors in multi-armed bandits

    Akram Erraqabi, Alessandro Lazaric, Michal Valko, Emma Brunskill, and Yun-En Liu. Trading off rewards and errors in multi-armed bandits. In Artificial Intelligence and Statistics, pages 709--717. PMLR, 2017

    2017

  16. [24]

    Causal inference in network experiments: regression-based analysis and design-based properties

    Mengsi Gao and Peng Ding. Causal inference in network experiments: regression-based analysis and design-based properties. arXiv preprint arXiv:2309.07476, 2023

    2023 arXiv

  17. [25]

    Confidence intervals for policy evaluation in adaptive experiments

    Vitor Hadad, David A Hirshberg, Ruohan Zhan, Stefan Wager, and Susan Athey. Confidence intervals for policy evaluation in adaptive experiments. Proceedings of the national academy of sciences, 118 0 (15): 0 e2014602118, 2021

    2021

  18. [26]

    Online statistical inference for matrix contextual bandit

    Qiyu Han, Will Wei Sun, and Yichen Zhang. Online statistical inference for matrix contextual bandit. arXiv preprint arXiv:2212.11385, 2022

    2022 arXiv

  19. [27]

    A simple and provably efficient algorithm for asynchronous federated contextual linear bandits

    Jiafan He, Tianhao Wang, Yifei Min, and Quanquan Gu. A simple and provably efficient algorithm for asynchronous federated contextual linear bandits. arXiv preprint arXiv:2207.03106, 2022

    2022 arXiv

  20. [28]

    Karnin, Tomer Koren, Ronny Lempel, and Oren Somekh

    Eshcar Hillel, Zohar S. Karnin, Tomer Koren, Ronny Lempel, and Oren Somekh. Distributed exploration in multi-armed bandits. In Advances in Neural Information Processing Systems, 2013

    2013

  21. [29]

    Causal inference in the social sciences

    Guido W Imbens. Causal inference in the social sciences. Annual Review of Statistics and Its Application, 11, 2024

    2024

  22. [30]

    Jamieson, Matthew Malloy, Robert D

    Kevin G. Jamieson, Matthew Malloy, Robert D. Nowak, and S \'e bastien Bubeck. lil' ucb : An optimal exploration algorithm for multi-armed bandits. arXiv preprint arXiv:1312.7308, 2013

    2013 arXiv

  23. [31]

    Short-lived high-volume bandits

    Su Jia, Nishant Oli, Ian Anderson, Paul Duff, Andrew A Li, and Ramamoorthi Ravi. Short-lived high-volume bandits. In International Conference on Machine Learning, pages 14902--14929. PMLR, 2023

    2023

  24. [32]

    Multi-armed bandits with interference

    Su Jia, Peter Frazier, and Nathan Kallus. Multi-armed bandits with interference. arXiv preprint arXiv:2402.01845, 2024

    2024 arXiv

  25. [33]

    Optimal regret analysis of thompson sampling in stochastic multi-armed bandit problem with multiple plays

    Junpei Komiyama, Junya Honda, and Hiroshi Nakagawa. Optimal regret analysis of thompson sampling in stochastic multi-armed bandit problem with multiple plays. In International Conference on Machine Learning, pages 1152--1161. PMLR, 2015

    2015

  26. [34]

    Position-based multiple-play bandit problem with unknown position bias

    Junpei Komiyama, Junya Honda, and Akiko Takeda. Position-based multiple-play bandit problem with unknown position bias. Advances in Neural Information Processing Systems, 30, 2017

    2017

  27. [35]

    Combinatorial cascading bandits

    Branislav Kveton, Zheng Wen, Azin Ashkan, and Csaba Szepesvari. Combinatorial cascading bandits. Advances in Neural Information Processing Systems, 28, 2015

    2015

  28. [36]

    Multiple-play bandits in the position-based model

    Paul Lagr \'e e, Claire Vernade, and Olivier Cappe. Multiple-play bandits in the position-based model. Advances in Neural Information Processing Systems, 29, 2016

    2016

  29. [37]

    Bandit algorithms

    Tor Lattimore and Csaba Szepesv \'a ri. Bandit algorithms. Cambridge University Press, 2020

    2020

  30. [38]

    Causal inference under approximate neighborhood interference

    Michael P Leung. Causal inference under approximate neighborhood interference. Econometrica, 90 0 (1): 0 267--293, 2022 a

    2022

  31. [39]

    Rate-optimal cluster-randomized designs for spatial interference

    Michael P Leung. Rate-optimal cluster-randomized designs for spatial interference. The Annals of Statistics, 50 0 (5): 0 3064--3087, 2022 b

    2022

  32. [40]

    Network cluster-robust inference

    Michael P Leung. Network cluster-robust inference. Econometrica, 91 0 (2): 0 641--667, 2023

    2023

  33. [41]

    Communication efficient federated learning for generalized linear bandits

    Chuanhao Li and Hongning Wang. Communication efficient federated learning for generalized linear bandits. arXiv preprint arXiv:2202.01087, 2022

    2022 arXiv

  34. [42]

    Contextual combinatorial cascading bandits

    Shuai Li, Baoxiang Wang, Shengyu Zhang, and Wei Chen. Contextual combinatorial cascading bandits. In International conference on machine learning, pages 1245--1253. PMLR, 2016

    2016

  35. [43]

    An experimental design for anytime-valid causal inference on multi-armed bandits

    Biyonka Liang and Iavor Bojinov. An experimental design for anytime-valid causal inference on multi-armed bandits. arXiv preprint arXiv:2311.05794, 2023

    2023 arXiv

  36. [44]

    A multiple-play bandit algorithm applied to recommender systems

    Jonathan Lou \"e dec, Max Chevalier, Josiane Mothe, Aur \'e lien Garivier, and S \'e bastien Gerchinovitz. A multiple-play bandit algorithm applied to recommender systems. In The Twenty-Eighth International Flairs Conference, 2015

    2015

  37. [45]

    Statistical inference for the mean outcome under a possibly non-unique optimal treatment strategy

    Alexander R Luedtke and Mark J Van Der Laan. Statistical inference for the mean outcome under a possibly non-unique optimal treatment strategy. Annals of statistics, 44 0 (2): 0 713, 2016

    2016

  38. [46]

    Managing the covid-19 pandemic crisis and changing welfare regimes, 2021

    Ka Ho Mok, Yeun-Wen Ku, and Tauchid Komara Yuda. Managing the covid-19 pandemic crisis and changing welfare regimes, 2021

    2021

  39. [47]

    Treatment effects in market equilibrium

    Evan Munro, Stefan Wager, and Kuang Xu. Treatment effects in market equilibrium. arXiv preprint arXiv:2109.11647, 2021

    2021 arXiv

  40. [48]

    Changing climates of conflict: A social network experiment in 56 schools

    Elizabeth Levy Paluck, Hana Shepherd, and Peter M Aronow. Changing climates of conflict: A social network experiment in 56 schools. Proceedings of the National Academy of Sciences, 113 0 (3): 0 566--571, 2016

    2016

  41. [49]

    Contextual combinatorial bandit and its application on diversified online recommendation

    Lijing Qin, Shouyuan Chen, and Xiaoyan Zhu. Contextual combinatorial bandit and its application on diversified online recommendation. In Proceedings of the 2014 SIAM International Conference on Data Mining, pages 461--469. SIAM, 2014

    2014

  42. [50]

    Donald B. Rubin. Randomization analysis of experimental data: The fisher randomization test comment. Journal of the American Statistical Association, 75: 0 591--593, 1980

    1980

  43. [51]

    Causal inference using potential outcomes: Design, modeling, decisions

    Donald B Rubin. Causal inference using potential outcomes: Design, modeling, decisions. Journal of the American Statistical Association, 100 0 (469): 0 322--331, 2005

    2005

  44. [52]

    Combinatorial bandits with relative feedback

    Aadirupa Saha and Aditya Gopalan. Combinatorial bandits with relative feedback. Advances in Neural Information Processing Systems, 32, 2019

    2019

  45. [53]

    Multi-armed bandit experimental design: Online decision-making and adaptive inference

    David Simchi-Levi and Chonghuan Wang. Multi-armed bandit experimental design: Online decision-making and adaptive inference. Management Science, 2024. doi:10.1287/mnsc.2023.00492

    2024

  46. [54]

    o r \'e nyi, R \'o bert Busa-Fekete, Istv \'a n Heged \

    Bal \'a zs Sz \"o r \'e nyi, R \'o bert Busa-Fekete, Istv \'a n Heged \"u s, R \'o bert Orm \'a ndi, M \'a rk Jelasity, and Bal \'a zs K \'e gl. Gossip-based distributed stochastic bandit algorithms. In International Conference on Machine Learning, 2013

    2013

  47. [55]

    Algorithms for adversarial bandit problems with multiple plays

    Taishi Uchiya, Atsuyoshi Nakamura, and Mineichi Kudo. Algorithms for adversarial bandit problems with multiple plays. In International Conference on Algorithmic Learning Theory, pages 375--389. Springer, 2010

    2010

  48. [56]

    Causal clustering: design of cluster experiments under network interference

    Davide Viviano, Lihua Lei, Guido Imbens, Brian Karrer, Okke Schrijvers, and Liang Shi. Causal clustering: design of cluster experiments under network interference. arXiv preprint arXiv:2310.14983, 2023

    2023 arXiv

  49. [57]

    Distributed bandit learning: Near-optimal regret with efficient communication

    Yuanhao Wang, Jiachen Hu, Xiaoyu Chen, and Liwei Wang. Distributed bandit learning: Near-optimal regret with efficient communication. arXiv preprint arxiv: 1904.06309, 2019

    1904 arXiv

  50. [58]

    Contextual bandits in a collaborative environment

    Qingyun Wu, Huazheng Wang, Quanquan Gu, and Hongning Wang. Contextual bandits in a collaborative environment. Proceedings of the 39th International ACM SIGIR conference on Research and Development in Information Retrieval, 2016

    2016

  51. [59]

    Linear contextual bandits with interference

    Yang Xu, Wenbin Lu, and Rui Song. Linear contextual bandits with interference. arXiv preprint arXiv:2409.15682, 2024

    2024 arXiv

  52. [60]

    A framework for multi-a (rmed)/b (andit) testing with online fdr control

    Fanny Yang, Aaditya Ramdas, Kevin G Jamieson, and Martin J Wainwright. A framework for multi-a (rmed)/b (andit) testing with online fdr control. Advances in Neural Information Processing Systems, 30, 2017

    2017

  53. [61]

    Power constrained bandits

    Jiayu Yao, Emma Brunskill, Weiwei Pan, Susan Murphy, and Finale Doshi-Velez. Power constrained bandits. In Machine Learning for Healthcare Conference, pages 209--259. PMLR, 2021

    2021

  54. [62]

    Inference for batched bandits

    Kelly Zhang, Lucas Janson, and Susan Murphy. Inference for batched bandits. Advances in Neural Information Processing Systems, 33: 0 9818--9829, 2020

    2020

  55. [63]

    Statistical inference with m-estimators on adaptively collected data

    Kelly Zhang, Lucas Janson, and Susan Murphy. Statistical inference with m-estimators on adaptively collected data. Advances in Neural Information Processing Systems, 34: 0 7460--7471, 2021

    2021

  56. [64]

    Individualized policy evaluation and learning under clustered network interference

    Yi Zhang and Kosuke Imai. Individualized policy evaluation and learning under clustered network interference. arXiv preprint arXiv:2311.02467, 2023

    2023 arXiv

  57. [65]

    Budget-constrained multi-armed bandits with multiple plays

    Datong Zhou and Claire Tomlin. Budget-constrained multi-armed bandits with multiple plays. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 32, 2018

    2018

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.