Pith. sign in

REVIEW 3 major objections 6 minor 65 references

Data Acquisition for Improving Model Fairness using Reinforcement Learning

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read The paper claims that a multi-armed bandit that ranks candidate data points by their predicted effect on demographic parity can shrink a model's fairness gap much faster than random, entropy, or accuracy-focused acquisition.

desk verdict Solid idea, but the written algorithm has a reward-sign bug that undermines the central claim; fixable and worth a look. read the letter →

arxiv 2412.03009 v1 pith:25TDA4GG submitted 2024-12-04 cs.LG cs.CY

classification cs.LGcs.CY
keywords dataacquisitionalgorithmicfairnessdemographicparitymulti-armedbanditsupperconfidenceboundinfluencefunctionsvaluationreinforcementlearning
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

This paper aims to show that model fairness can be improved at the data-procurement stage: instead of buying arbitrary or balanced extra labels, an algorithm can decide which specific points from a pool are worth acquiring. It proposes DataSift, which splits the pool into partitions, treats each partition as an arm in a multi-armed bandit, and uses a reward score that rewards partitions whose data improves demographic parity while preserving accuracy. A faster variant, DataSift-Inf, ranks points inside the chosen partition by the influence functions predict each would have on fairness, so each batch is deliberately constructed rather than random. Across six datasets, the authors report that even a small acquisition budget reduces the demographic-parity gap substantially, often matching or beating acquisition of the entire pool.

What carries the argument

The load-bearing pieces are the UCB bandit over partitions and the influence-function valuation layer. For DataSift, Equation (1) is the reward-spreading rule: when partition $C_i$ is evaluated, partition $C_j$ receives score $r_j = \Delta F / ((1 + |\Delta BR_{C_j}|)(1 + \mathrm{dist}(C_i,C_j)))$, where $\Delta F$ is the observed demographic-parity change, $\Delta BR$ is the intra-partition base-rate difference, and $\mathrm{dist}$ is normalized centroid distance. For DataSift-Inf, the valuation machinery is the first-order influence function $I_\theta(d) = -H_\theta^{-1} \nabla_\theta L(d,\theta^*)$, applied to fairness $F$ through the chain rule, then transferred to unlabeled pool points by a ridge regressor $R$ trained on (training point, predicted influence) pairs. The regressor's output orders each partition, and the top-$K$ points form each batch.

What would settle it

Take any of the six datasets and a random sample of unseen pool points; for each, retrain once with that single point added, record the actual change in demographic parity, and correlate those changes with the influences predicted by the ridge regressor. A near-zero or negative correlation would falsify the valuation layer, and then DataSift-Inf would reduce to random batch selection.

Watch

Extended reading notes

Core claim

On the authors' own terms, the central discovery is that fairness is a property that can be bought selectively: the improvement in demographic parity depends less on how many points are added than on which points are added, and the right ordering can be learned. DataSift formalizes acquisition as a UCB bandit problem where the arms are data-pool partitions and the reward for a partition is the measured fairness improvement of a batch, shared with other partitions according to their base-rate differences and centroid distances. DataSift-Inf adds a valuation layer: first-order influence functions computed on the current training model estimate each training point's effect on the fairness metric, a ridge regressor learns to predict those influences for points in the pool, and the top-ranked points in the selected partition form the next batch. The reported result is that DataSift-Inf reaches the fairness threshold with less than half the budget in most datasets and stops early, while the plain bandit already outperforms random, entropy, and accuracy-oriented baselines.

Load-bearing premise

Everything rests on the assumption that the influence of a training point on fairness, computed with the current model, still predicts what happens when that point is actually added, and that a regressor can transfer that prediction to points in the pool; on top of that, DataSift's reward-sharing formula in Equation (1) is a heuristic rather than a derived fairness model.

Editorial extensions

If this is right

  • Acquiring a small, well-chosen set of additional labels can close most of the demographic-parity gap, in some cases matching or beating a model trained on the entire data pool.
  • Balanced acquisition across demographic groups is not a reliable fairness strategy; selective ordering is what carries the benefit.
  • DataSift-Inf can stop early once fairness crosses a threshold, so the labeling budget and retraining cost both shrink.
  • The approach preserves or improves accuracy on most tested datasets even though it optimizes for fairness, with slight drops on two datasets that a larger budget can offset.

Reading between the lines

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

  • The valuation-plus-bandit design is not tied to demographic parity: any fairness signal that can be computed after each batch, or differentiated where needed, could be plugged into Equation (1), so equalized odds or calibration gaps are natural next targets.
  • Because DataSift-Inf's regressor is trained on the current model's influence values, its reliability should degrade as the model moves far from the point where those influences were computed; a direct test would measure correlation between predicted and actual per-point fairness effects on held-out pool points.
  • A fixed early-stopping threshold caused oscillation in at least one dataset; a dataset-adaptive threshold derived from the initial disparity would make the budget-to-fairness trade-off more predictable.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper studies the problem of acquiring additional labeled data to improve the demographic parity fairness of a downstream classifier. It proposes DATASIFT, an MAB-based approach that partitions the data pool, treats partitions as arms, and uses a UCB policy with a fairness-aware reward that also incorporates base-rate difference and inter-partition distance. A second method, DATASIFT-Inf, augments this with influence functions to rank data points within a partition and build batches without retraining. The authors empirically compare both methods against Random, Entropy, Inf, and AutoData on six real-world datasets and several classifiers, reporting that DATASIFT improves fairness faster and with less budget than baselines, and that DATASIFT-Inf further accelerates convergence.

Significance. If the results hold, the paper addresses a timely and underexplored problem: how to prioritize additional data acquisition to mitigate group fairness violations, rather than applying post-hoc bias mitigation. The MAB formulation and the integration of influence functions for fairness-oriented data valuation are reasonable and could be of practical value. The authors provide a link to source code, which is a positive step for reproducibility. However, the paper's central mechanism contains a sign inconsistency between the reward and the acceptance criterion, the UCB update is undefined for arms without positive rewards, and the empirical evaluation lacks variance information. These issues currently undermine the validity of the reported gains and need to be resolved before the contribution can be assessed reliably.

major comments (3)
  1. [3.2, Eq. (1) and Algorithm 1 lines 12–13] The reward definition in Eq. (1) is inconsistent with the acceptance criterion in Algorithm 1. The algorithm retains a batch when |F(D^train ∪ b)| ≤ |F_best| (line 13), i.e., when the absolute demographic parity gap decreases. However, Eq. (1) rewards partitions with r_j = ΔF / ((1+|ΔBR_Cj|)(1+dist(C_i,C_j))) where ΔF = F_{new} − F_{old}, and the text states that ΔF > 0 is a reward and ΔF < 0 is a penalty. When the current model has F > 0 (bias against the privileged group), moving toward zero gives ΔF < 0, so the very partitions that improve fairness are penalized; the same occurs when an acquisition overshoots across zero. Since Figure 5 plots signed F values (negative and positive), the precondition F < 0 throughout is not met. This sign error means the described UCB mechanism would push the search away from useful partitions, and the reported fairness gains cannot be produced by the algorithm as written. The paper must correct Eq. (1) to reward signed absolute improvement, e.g., r_j = −(|F_new| − |F_old|)/... or an equivalent, and re-run all experiments.
  2. [3.2, definition of n_i^k and Eq. (2)] The definition of n_i^k as the number of times C_i is 'selected and rewarded a positive score' breaks the UCB update. R_i^k = (1/n_i^k) Σ r_j^i is undefined for any arm that has never received a positive reward (n_i^k = 0). Moreover, if no arm has a positive reward after some iterations, n^k = Σ_i n_i^k = 0, making ln(n^k) in Eq. (2) undefined. This is not a corner case: the sign error in Eq. (1) makes negative rewards common, so the algorithm cannot compute U_i^k for the arms it most needs to explore. The paper should redefine n_i^k as the number of times the arm was selected (regardless of reward), and define R_i^k as the average over all observed rewards (or use a standard UCB with a prior).
  3. [4, Figures 3–6 and Table 2] The central empirical claim that DATASIFT 'consistently outperforms' baselines is supported only by single-run curves. The methods involve random batch sampling (DataSift), stochastic partition selection (GMM), and MAB exploration, so the reported trajectories are subject to high variance. Without error bars, standard deviations over multiple seeds, or significance tests, the reader cannot determine whether the observed advantages are real or artifacts of one run. Please report means and standard deviations over at least 5–10 random seeds (or equivalent) for all main figures, and perform a paired significance test where appropriate.
minor comments (6)
  1. [4.3] There is a typo: 'Alorithm 2' should be 'Algorithm 2'.
  2. [3.3.2] 'know to be NP-hard' should be 'known to be NP-hard'.
  3. [3.2] 'to acquire mode data points' should be 'to acquire more data points'.
  4. [Figures 3 and 5] The y-axis label is 'Fairness Metric' in both, but Figure 3 uses absolute disparity (non-negative) whereas Figure 5 shows signed F values (including negative numbers). Please use a consistent definition, e.g., plot |F| or label the signed quantity clearly.
  5. [Algorithm 2 and Section 3.3.2] Algorithm 2 line 3 says 'Train Regressor R' but does not specify the model class; later text refers to a ridge regressor. Please state the regressor in the algorithm.
  6. [4.1.3] The source code link appears as the word 'DATASIFT' without a URL; include a working hyperlink or footnote.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper is an empirical algorithm evaluation whose reported fairness gains are measured by retraining, not by its fitted or heuristic components.

full rationale

DataSift and DataSift-Inf are evaluated by acquiring data, retraining the downstream model, and measuring demographic parity on a held-out test set; the reported improvement is not constructed from the reward formula or the influence regressor. The influence-function regressor in Section 3.3.2 is fitted to training-point influences and used only to rank pool points; the final fairness numbers come from actual retraining, so this is a heuristic rather than a fitted input renamed as a prediction. Equation (1)'s reward uses the change in the same fairness metric that appears in the problem definition, but that is objective alignment (the algorithm is designed to maximize fairness improvement), not definitional equivalence. The sole self-citation by an overlapping author, reference [48], appears only in the related-work survey of data valuation and is not load-bearing for any derivation. Potential concerns such as the sign convention in Equation (1) relative to the acceptance condition in Algorithm 1 and the undefined UCB count when n_i^k = 0 are correctness or robustness issues, not circularity: they do not make the output equal to the input by construction.

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

The method relies on several hand-chosen hyperparameters and heuristic assumptions. The core contributions are empirical, so the ledger is modest: four tuned quantities, three domain assumptions, and a new reward-spreading heuristic. No new physical or theoretical entities are introduced.

free parameters (4)
  • alpha (exploration-exploitation trade-off) = 0.1
    Chosen by hand in Section 4.1.3; ablation (Figure 7b) shows sensitivity.
  • fairness threshold tau = 0.01
    Fixed early stopping threshold in Section 4.1.3; the paper notes it can cause excessive iterations on some datasets (Section 4.5).
  • batch size K = 10% of budget
    Set as percentage of budget in Section 4.1.3; ablation (Figure 7a) shows best range 6-12%.
  • number of partitions g = determined by BIC on GMM
    Partition count chosen via BIC in Section 4.1.3; it affects arm granularity and results.
assumptions (4)
  • domain assumption Loss function is twice differentiable and strictly convex so Hessian inverse exists
    Section 3.3.1; limits influence functions to parametric models like logistic regression and SVM.
  • ad hoc to paper Reward can be spread to other partitions based on base rate difference and centroid distance
    Equation (1) assumes partitions with similar base rates and distances should receive similar rewards, a heuristic introduced without derivation.
  • domain assumption All data points have uniform acquisition cost and data pool is homogeneous
    Section 2 simplifies the problem by ignoring differential costs and multiple sources.
  • ad hoc to paper A regressor trained on training-data influences predicts influences of pool points
    Section 3.3.2 uses a ridge regression model to transfer influence estimates to pool points; no theoretical guarantee is given.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Data Acquisition for Improving Model Fairness using Reinforcement Learning." pith.science (2026). https://pith.science/paper/25TDA4GG

@misc{pith2026241203009,
  author       = {Pith},
  title        = {Pith review of: Data Acquisition for Improving Model Fairness using Reinforcement Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/25TDA4GG}},
  note         = {Machine review of arXiv:2412.03009}
}
read the original abstract

Machine learning systems are increasingly being used in critical decision making such as healthcare, finance, and criminal justice. Concerns around their fairness have resulted in several bias mitigation techniques that emphasize the need for high-quality data to ensure fairer decisions. However, the role of earlier stages of machine learning pipelines in mitigating model bias has not been explored well. In this paper, we focus on the task of acquiring additional labeled data points for training the downstream machine learning model to rapidly improve its fairness. Since not all data points in a data pool are equally beneficial to the task of fairness, we generate an ordering in which data points should be acquired. We present DataSift, a data acquisition framework based on the idea of data valuation that relies on partitioning and multi-armed bandits to determine the most valuable data points to acquire. Over several iterations, DataSift selects a partition and randomly samples a batch of data points from the selected partition, evaluates the benefit of acquiring the batch on model fairness, and updates the utility of partitions depending on the benefit. To further improve the effectiveness and efficiency of evaluating batches, we leverage influence functions that estimate the effect of acquiring a batch without retraining the model. We empirically evaluate DataSift on several real-world and synthetic datasets and show that the fairness of a machine learning model can be significantly improved even while acquiring a few data points.

Figures

Figures reproduced from arXiv: 2412.03009 by the authors.

Figure 1
Figure 1. Example showing that acquiring the same number of data points across demographic groups does no better [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Overview of DATASIFT for data acquisition to improve model fairness. The blue-shaded part highlights DATASIFT with multi-armed bandit (MAB) framework while the red-shaded part incorporates data valuation with the MAB framework. most valuable data points with an O(N × T r) complexity. Note that data are often evaluated in batches rather than one data point at a time for efficiency purposes. While cheaper than the bru… view at source ↗
Figure 3
Figure 3. Comparing DATASIFT with baselines (AutoData, Random, and Entropy) to highlight the effectiveness in achieving fairness. DATASIFT consistently outperforms the other methods in improving fairness for most of the datasets. The black dotted line indicates ultimate fairness if the entire data pool is acquired (i.e., FDtrain∪D). of the budget, and the target fairness threshold τ was set at 0.01. The trade-off parameter be… view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Generalizability of DATASIFT to black-box ML algorithms (ACSIncome dataset, Black dotted line: FDtrain∪D) 10 [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 5
Figure 5. Figure 5: Comparing DATASIFT-Inf, our MAB-approach based on data valuation, with DATASIFT and Inf to highlight the importance of data valuation in acquiring data that rapidly improve model fairness. DATASIFT-Inf exhibits the most improvement in fairness with the least amount of …
Figure 6
Figure 6. Figure 6: Evaluation of data valuation across parametric models ( [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 7
Figure 7. Figure 7: Effect of Hyper-parameters and design choices (ACSIncome dataset, DS: D [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]
Figure 8
Figure 8. Figure 8: Scalability on synthetic dataset. Using the AdultIncome dataset, we synthetically generate additional data points to control the dataset size, which ranges from the original size up to 106 data points. Throughout this study, we maintain a fixed budget and corresponding…
Figure 9
Figure 9. Figure 9: Efficiency on real-world datasets. requirement across all cases. In contrast, the batch selection complexity for DATASIFT is influenced by partition size, leading to a slight upward trend in efficiency, which is nearly negligible. 5 Related Work The study in this paper…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

65 extracted references · 48 canonical work pages

  1. [1]

    Analysis of thompson sampling for the multi-armed bandit problem

    Shipra Agrawal and Navin Goyal. Analysis of thompson sampling for the multi-armed bandit problem. In Conference on Learning Theory , pages 39–1, Edinburgh, Scotland, 2012. JMLR Workshop and Conference Proceedings

  2. [2]

    Machine bias: There’s software used across the country to predict future criminals

    Julia Angwin, Jeff Larson, Surya Mattu, and Lauren Kirchner. Machine bias: There’s software used across the country to predict future criminals. and it’s biased against blacks., 2016. ProPublica, May 23, 2016

  3. [3]

    Abolfazl Asudeh, Zhongjun Jin, and H. V . Jagadish. Assessing and Remedying Coverage for a Given Dataset. In 2019 IEEE 35th International Conference on Data Engineering (ICDE), pages 554–565, Macau, China, April

  4. [4]

    Using upper confidence bounds for online learning

    Peter Auer. Using upper confidence bounds for online learning. In Proceedings of the 41st Annual Symposium on Foundations of Computer Science (FOCS), pages 270–279, Redondo Beach, CA, USA, 2000. IEEE, IEEE

  5. [5]

    Using confidence bounds for exploitation-exploration trade-offs

    Peter Auer. Using confidence bounds for exploitation-exploration trade-offs. Journal of Machine Learning Research, 3(Nov):397–422, 2002

  6. [6]

    Learn2clean: Optimizing the sequence of tasks for web data preparation

    Laure Berti-Equille. Learn2clean: Optimizing the sequence of tasks for web data preparation. In The world wide web conference, pages 2580–2586, 2019

  7. [7]

    The shapley value in database manage- ment

    Leopoldo Bertossi, Benny Kimelfeld, Ester Livshits, and Mikaël Monet. The shapley value in database manage- ment. ACM Sigmod Record, 52(2):6–17, 2023

  8. [8]

    Fair preprocessing: towards understanding compositional fairness of data transformers in machine learning pipeline

    Sumon Biswas and Hridesh Rajan. Fair preprocessing: towards understanding compositional fairness of data transformers in machine learning pipeline. In Proceedings of the 29th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering, page 981–993, 2021

Show all 65 references
  1. [9]

    The art and practice of data science pipelines: A comprehensive study of data science pipelines in theory, in-the-small, and in-the-large

    Sumon Biswas, Mohammad Wardat, and Hridesh Rajan. The art and practice of data science pipelines: A comprehensive study of data science pipelines in theory, in-the-small, and in-the-large. In Proceedings of the 44th International Conference on Software Engineering, page 2091–2...

  2. [10]

    Housing department slaps facebook with discrimination charge

    Brakkton Booker. Housing department slaps facebook with discrimination charge. lhttps://www.npr.org/ 2019/03/28/707614254/hud-slaps-facebook-with-housing-discrimination-charge , 2019

  3. [11]

    Bandits with heavy tail

    Sébastien Bubeck, Nicolo Cesa-Bianchi, and Gábor Lugosi. Bandits with heavy tail. IEEE Transactions on Information Theory, 59(11):7711–7717, 2013

  4. [12]

    A survey on deep reinforcement learning for data processing and analytics

    Qingpeng Cai, Can Cui, Yiyuan Xiong, Wei Wang, Zhongle Xie, and Meihui Zhang. A survey on deep reinforcement learning for data processing and analytics. IEEE Transactions on Knowledge and Data Engineering, 35(5):4446–4465, 2023

  5. [13]

    Upper- confidence-bound algorithms for active learning in multi-armed bandits

    Alexandra Carpentier, Alessandro Lazaric, Mohammad Ghavamzadeh, Rémi Munos, and Peter Auer. Upper- confidence-bound algorithms for active learning in multi-armed bandits. In International Conference on Algorith- mic Learning Theory, pages 189–203. Springer, 2011

  6. [14]

    Fairness in machine learning: A survey

    Simon Caton and Christian Haas. Fairness in machine learning: A survey. ACM Comput. Surv., 56(7), April 2024

  7. [15]

    Selective data acquisition in the wild for model charging

    Chengliang Chai, Jiabin Liu, Nan Tang, Guoliang Li, and Yuyu Luo. Selective data acquisition in the wild for model charging. Proc. VLDB Endow., 15(7):1466–1478, March 2022

  8. [16]

    Fair prediction with disparate impact: A study of bias in recidivism prediction instruments

    Alexandra Chouldechova. Fair prediction with disparate impact: A study of bias in recidivism prediction instruments. CoRR, abs/1703.00056, 2017

  9. [17]

    Dennis Cook and Sanford Weisberg

    R. Dennis Cook and Sanford Weisberg. Characterizations of an empirical influence function for detecting influential cases in regression. Technometrics, 22(4):495–508, 1980

  10. [18]

    Give me some credit, 2011

    Will Cukierski Credit Fusion. Give me some credit, 2011

  11. [19]

    Rpt-insight-amazon scraps secret ai recruiting tool that showed bias against women

    Jeffrey Dastin. Rpt-insight-amazon scraps secret ai recruiting tool that showed bias against women. Reuters, 2018

  12. [20]

    Davis, S

    G. Davis, S. Mallat, and M. Avellaneda. Adaptive greedy approximations. Constructive Approximation, 13(1):57, 1997. 16 Data Acquisition For Improving Fairness Using Reinforcement Learning

  13. [21]

    Explanations for data repair through shapley values

    Daniel Deutch, Nave Frost, Amir Gilad, and Oren Sheffer. Explanations for data repair through shapley values. In Proceedings of the 30th ACM International Conference on Information & Knowledge Management, CIKM ’21, page 362–371, New York, NY , USA, 2021. Association for Comput...

  14. [22]

    Retiring adult: New datasets for fair machine learning

    Frances Ding, Moritz Hardt, John Miller, and Ludwig Schmidt. Retiring adult: New datasets for fair machine learning. Advances in neural information processing systems, 34:6478–6490, 2021

  15. [23]

    Uci machine learning repository

    Dheeru Dua, Casey Graff, et al. Uci machine learning repository. 2017

  16. [24]

    A density-based algorithm for discovering clusters in large spatial databases with noise

    Martin Ester, Hans-Peter Kriegel, Jörg Sander, Xiaowei Xu, et al. A density-based algorithm for discovering clusters in large spatial databases with noise. In kdd, volume 96, pages 226–231, 1996

  17. [25]

    Joint entity linking with deep reinforcement learning

    Zheng Fang, Yanan Cao, Qian Li, Dongjie Zhang, Zhenyu Zhang, and Yanbing Liu. Joint entity linking with deep reinforcement learning. In The World Wide Web Conference, WWW ’19, page 438–447, New York, NY , USA,

  18. [26]

    Mario A. T. Figueiredo and Anil K. Jain. Unsupervised learning of finite mixture models. IEEE Transactions on pattern analysis and machine intelligence, 24(3):381–396, 2002

  19. [27]

    Association for Computing Machinery

  20. [28]

    Data shapley: Equitable valuation of data for machine learning

    Amirata Ghorbani and James Zou. Data shapley: Equitable valuation of data for machine learning. In Proceedings of the 36th International Conference on Machine Learning, volume 97, pages 2242–2251. PMLR, 09–15 Jun 2019

  21. [29]

    On upper-confidence bound policies for switching bandit problems

    Aurélien Garivier and Eric Moulines. On upper-confidence bound policies for switching bandit problems. In International conference on algorithmic learning theory, pages 174–188. Springer, 2011

  22. [30]

    Deepline: Automl tool for pipelines generation using deep reinforcement learning and hierarchical actions filtering

    Yuval Heffetz, Roman Vainshtein, Gilad Katz, and Lior Rokach. Deepline: Automl tool for pipelines generation using deep reinforcement learning and hierarchical actions filtering. In Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Minin...

  23. [31]

    Self-driving cars more likely to hit blacks

    Karen Hao. Self-driving cars more likely to hit blacks. lhttps://www.technologyreview.com/2019/03/ 01/136808/self-driving-cars-are-coming-but-accidents-may-not-be-evenly-distributed/ , 2019

  24. [32]

    Data clustering: 50 years beyond k-means

    Anil K Jain. Data clustering: 50 years beyond k-means. Pattern recognition letters, 31(8):651–666, 2010

  25. [33]

    Kenneth Holstein, Jennifer Wortman Vaughan, Hal Daumé, Miro Dudik, and Hanna Wallach. Improving fairness in machine learning systems: What do industry practitioners need? In Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems, page 1–16, 2019

  26. [34]

    Understanding black-box predictions via influence functions

    Pang Wei Koh and Percy Liang. Understanding black-box predictions via influence functions. In Proceedings of the 34th International Conference on Machine Learning, volume 70, pages 1885–1894, 06–11 Aug 2017

  27. [35]

    Matthew Kay, Cynthia Matuszek, and Sean A. Munson. Unequal representation and gender stereotypes in image search results for occupations. In Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems, page 3819–3828, 2015

  28. [36]

    Qtune: a query-aware database tuning system with deep reinforcement learning

    Guoliang Li, Xuanhe Zhou, Shifu Li, and Bo Gao. Qtune: a query-aware database tuning system with deep reinforcement learning. Proc. VLDB Endow., 12(12):2118–2130, August 2019

  29. [37]

    Algorithms for multi-armed bandit problems

    V olodymyr Kuleshov and Doina Precup. Algorithms for multi-armed bandit problems. arXiv preprint arXiv:1402.6028, 2014

  30. [38]

    Data acquisition for improving model confidence

    Yifan Li, Xiaohui Yu, and Nick Koudas. Data acquisition for improving model confidence. Proc. ACM Manag. Data, 2(3), May 2024

  31. [39]

    Data Acquisition for Improving Machine Learning Models

    Yifan Li, Xiaohui Yu, and Nick Koudas. Data Acquisition for Improving Machine Learning Models. Proceedings of the VLDB Endowment, 14(10):1832–1844, June 2021. arXiv:2105.14107 [cs]

  32. [40]

    A Survey on Bias and Fairness in Machine Learning

    Ninareh Mehrabi, Fred Morstatter, Nripsuta Saxena, Kristina Lerman, and Aram Galstyan. A Survey on Bias and Fairness in Machine Learning. ACM Computing Surveys, 54(6):1–35, July 2022

  33. [41]

    Applications and computation of the shapley value in databases and machine learning

    Xuan Luo and Jian Pei. Applications and computation of the shapley value in databases and machine learning. In Companion of the 2024 International Conference on Management of Data, pages 630–635, 2024

  34. [42]

    Tailoring data source distributions for fairness-aware data integration

    Fatemeh Nargesian, Abolfazl Asudeh, and HV Jagadish. Tailoring data source distributions for fairness-aware data integration. Proceedings of the VLDB Endowment, 14(11):2519–2532, 2021. 17 Data Acquisition For Improving Fairness Using Reinforcement Learning

  35. [43]

    Improving information extraction by acquiring external evidence with reinforcement learning

    Karthik Narasimhan, Adam Yala, and Regina Barzilay. Improving information extraction by acquiring external evidence with reinforcement learning. In Jian Su, Kevin Duh, and Xavier Carreras, editors, Proceedings of the 2016 Conference on Empirical Methods in Natural Language Pro...

  36. [44]

    Recommendation system-based upper confidence bound for online advertising

    Nhan Nguyen-Thanh, Dana Marinca, Kinda Khawam, David Rohde, Flavian Vasile, Elena Simona Lohan, Steven Martin, and Dominique Quadri. Recommendation system-based upper confidence bound for online advertising. arXiv preprint arXiv:1909.04190, 2019

  37. [45]

    The bayesian information criterion: background, derivation, and applications

    Andrew A Neath and Joseph E Cavanaugh. The bayesian information criterion: background, derivation, and applications. Wiley Interdisciplinary Reviews: Computational Statistics, 4(2):199–203, 2012

  38. [46]

    Scikit-learn: Machine learning in python

    Fabian Pedregosa, Gaël Varoquaux, Alexandre Gramfort, Vincent Michel, Bertrand Thirion, Olivier Grisel, Mathieu Blondel, Peter Prettenhofer, Ron Weiss, Vincent Dubourg, et al. Scikit-learn: Machine learning in python. the Journal of machine Learning research, 12:2825–2830, 2011

  39. [47]

    Pytorch: An imperative style, high-performance deep learning library

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. Pytorch: An imperative style, high-performance deep learning library. Advances in neural information processing systems, 32, 2019

  40. [48]

    Interpretable Data-Based Explanations for Fairness Debugging

    Romila Pradhan, Jiongli Zhu, Boris Glavic, and Babak Salimi. Interpretable Data-Based Explanations for Fairness Debugging. In Proceedings of the 2022 International Conference on Management of Data, SIGMOD ’22, pages 247–261, New York, NY , USA, June 2022. Association for Compu...

  41. [49]

    What can data-centric AI learn from data and ML engineering? CoRR, abs/2112.06439, 2021

    Neoklis Polyzotis and Matei Zaharia. What can data-centric AI learn from data and ML engineering? CoRR, abs/2112.06439, 2021

  42. [50]

    Coverage-based data-centric approaches for responsible and trustworthy ai

    Nima Shahbazi, Mahdi Erfanian, and Abolfazl Asudeh. Coverage-based data-centric approaches for responsible and trustworthy ai. IEEE Data Engineering Bulletin, 2024

  43. [51]

    Sourcesight: Enabling effective source selection

    Theodoros Rekatsinas, Amol Deshpande, Xin Luna Dong, Lise Getoor, and Divesh Srivastava. Sourcesight: Enabling effective source selection. In Proceedings of the 2016 International Conference on Management of Data, pages 2157–2160, 2016

  44. [52]

    C. E. Shannon. A mathematical theory of communication. The Bell System Technical Journal, 27(3):379–423, 1948

  45. [53]

    Representation bias in data: A survey on identifica- tion and resolution techniques

    Nima Shahbazi, Yin Lin, Abolfazl Asudeh, and HV Jagadish. Representation bias in data: A survey on identifica- tion and resolution techniques. ACM Computing Surveys, 55(13s):1–39, 2023

  46. [54]

    Reinforcement learning: An introduction

    Richard S Sutton and Andrew G Barto. Reinforcement learning: An introduction. MIT press, 2018

  47. [55]

    Introduction to multi-armed bandits

    Aleksandrs Slivkins et al. Introduction to multi-armed bandits. Foundations and Trends® in Machine Learning, 12(1-2):1–286, 2019

  48. [56]

    Fairness definitions explained

    Sahil Verma and Julia Rubin. Fairness definitions explained. In Proceedings of the international workshop on software fairness, pages 1–7. Association for Computing Machinery, 2018

  49. [57]

    Slice tuner: A selective data acquisition framework for accurate and fair machine learning models

    Ki Hyun Tae and Steven Euijong Whang. Slice tuner: A selective data acquisition framework for accurate and fair machine learning models. In Proceedings of the 2021 International Conference on Management of Data, pages 1771–1783, 2021

  50. [58]

    Op- timizing data acquisition to enhance machine learning performance

    Tingting Wang, Shixun Huang, Zhifeng Bao, J Shane Culpepper, V olkan Dedeoglu, and Reza Arablouei. Op- timizing data acquisition to enhance machine learning performance. Proceedings of the VLDB Endowment , 17(6):1310–1323, 2024

  51. [59]

    Multi-armed bandit algorithms and empirical evaluation

    Joannes Vermorel and Mehryar Mohri. Multi-armed bandit algorithms and empirical evaluation. In European conference on machine learning, pages 437–448. Springer, 2005

  52. [60]

    Data collection and quality challenges in deep learning: a data-centric ai perspective

    Steven Euijong Whang, Yuji Roh, Hwanjun Song, and Jae-Gil Lee. Data collection and quality challenges in deep learning: a data-centric ai perspective. The VLDB Journal, 32(4):791–813, January 2023

  53. [61]

    Data discovery

    Gerhard Weikum. Data discovery. Data Science Journal, 12, 2013

  54. [62]

    Data-centric ai: Techniques and future perspectives

    Daochen Zha, Kwei-Herng Lai, Fan Yang, Na Zou, Huiji Gao, and Xia Hu. Data-centric ai: Techniques and future perspectives. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD ’23, page 5839–5840, New York, NY , USA, 2023. Association fo...

  55. [63]

    Data-centric AI: Perspectives and Challenges, pages 945–948

    Daochen Zha, Zaid Pervaiz Bhat, Kwei-Herng Lai, Fan Yang, and Xia Hu. Data-centric AI: Perspectives and Challenges, pages 945–948

  56. [65]

    An end-to-end automatic cloud database tuning system using deep reinforcement learning

    Ji Zhang, Yu Liu, Ke Zhou, Guoliang Li, Zhili Xiao, Bin Cheng, Jiashu Xing, Yangtao Wang, Tianheng Cheng, Li Liu, Minwei Ran, and Zekang Li. An end-to-end automatic cloud database tuning system using deep reinforcement learning. In Proceedings of the 2019 International Confere...

  57. [2019]

    ISSN: 2375-026X

    IEEE. ISSN: 2375-026X

Pith tools

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