1,475 Hits in 3.4 sec

Fairness in Learning: Classic and Contextual Bandits [article]

Matthew Joseph and Michael Kearns and Jamie Morgenstern and Aaron Roth
2016 arXiv   pre-print
fair contextual bandit algorithm, and conversely any fair contextual bandit algorithm can be transformed into a KWIK learning algorithm.  ...  First, in the important special case of the classic stochastic bandits problem (i.e., in which there are no contexts), we provide a provably fair algorithm based on "chained" confidence intervals, and  ...  2 (δ-fairness in the classic bandits setting).  ... 
arXiv:1605.07139v2 fatcat:esywof72wzcozkmv7yv74wefuy

Fairness in Machine Learning: A Survey [article]

Simon Caton, Christian Haas
2020 arXiv   pre-print
This article seeks to provide an overview of the different schools of thought and approaches to mitigating (social) biases and increase fairness in the Machine Learning literature.  ...  Although much of the literature emphasizes binary classification, a discussion of fairness in regression, recommender systems, unsupervised learning, and natural language processing is also provided along  ...  Adversarial Learning Beutel et al. [30, 28] Celis and Keswani [55] Edwards and Storkey [89] Feng et al. [97] Wadsworth et al. [271] Xu et al. [278] Zhang et al.  ... 
arXiv:2010.04053v1 fatcat:cvao3z5fmzc3xoehq5q7tqfvgy

Metric-Free Individual Fairness in Online Learning [article]

Yahav Bechavod, Christopher Jung, Zhiwei Steven Wu
2022 arXiv   pre-print
Surprisingly, in the stochastic setting where the data are drawn independently from a distribution, we are also able to establish PAC-style fairness and accuracy generalization guarantees (Rothblum and  ...  regret and number of fairness violations.  ...  in conjunction with the Israel National Cyber Directorate, and the Apple Scholars in AI/ML PhD Fellowship.  ... 
arXiv:2002.05474v6 fatcat:zyjxcdatunexld6qlvlcrq4aey

Calibrated Fairness in Bandits [article]

Yang Liu, Goran Radanovic, Christos Dimitrakakis, Debmalya Mandal, David C. Parkes
2017 arXiv   pre-print
We show that a variation on Thompson sampling satisfies smooth fairness for total variation distance, and give an Õ((kT)^2/3) bound on fairness regret.  ...  We study fairness within the stochastic, multi-armed bandit (MAB) decision making framework. We adapt the fairness framework of "treating similar individuals similarly" to this setting.  ...  [12] were the rst to introduce fairness concepts in the bandits se ing. ese authors adopt the notion of weak meritocratic fairness, and study it within the classic and contextual bandit se ing. eir  ... 
arXiv:1707.01875v1 fatcat:n5ucp45torbhvlvdo4apv5xxiy

Fairness in Learning-Based Sequential Decision Algorithms: A Survey [article]

Xueru Zhang, Mingyan Liu
2020 arXiv   pre-print
Algorithmic fairness in decision-making has been studied extensively in static settings where one-shot decisions are made on tasks such as classification.  ...  In each case the impact of various fairness interventions on the underlying population is examined.  ...  In a classic stochastic bandit problem, there is a set of arms Z = {1, · · · , K }.  ... 
arXiv:2001.04861v1 fatcat:itudnjgkuvckpglqgq32f2pjq4

Achieving User-Side Fairness in Contextual Bandits [article]

Wen Huang and Kevin Labille and Xintao Wu and Dongwon Lee and Neil Heffernan
2020 arXiv   pre-print
We formulate our fair personalized recommendation as a modified contextual bandit and focus on achieving fairness on the individual whom is being recommended an item as opposed to achieving fairness on  ...  We conduct numerous experimental evaluations to compare the performances of our fair contextual bandit to that of LinUCB and show that our approach achieves group-level fairness while maintaining a high  ...  The source code and datasets are available at Achieving User-Side Fairness in Contextual dl=0.  ... 
arXiv:2010.12102v1 fatcat:znuzhcuumfgjbipzreqovspvxu

Learning Fair Equilibria in Sponsored Search Auctions [article]

Andrea Celli, Riccardo Colini-Baldeschi, Stefano Leonardi
2021 arXiv   pre-print
In this work we investigate the strategic learning implications of the deployment of sponsored search auction mechanisms that obey to fairness criteria.  ...  We propose two mechanisms, β-Fair GSP and GSP-EFX, that compose GSP with, respectively, an envy-free up to one item (EF1), and an envy-free up to any item (EFX) fair division scheme.  ...  Fairness in online learning for ad auctions has also been studied within the the stochastic multi-armed bandit contextual setting.  ... 
arXiv:2107.08271v1 fatcat:vctfli55ercvlb2ne7ey3sww3y

Fairness of Exposure in Stochastic Bandits [article]

Lequn Wang, Yiwei Bai, Wen Sun, Thorsten Joachims
2021 arXiv   pre-print
We formulate fairness regret and reward regret in this setting, and present algorithms for both stochastic multi-armed bandits and stochastic linear bandits.  ...  Contextual bandit algorithms have become widely used for recommendation in online systems (e.g. marketplaces, music streaming, news), where they now wield substantial influence on which items get exposed  ...  Acknowledgements This research was supported in part by NSF Awards IIS-1901168 and IIS-2008139.  ... 
arXiv:2103.02735v2 fatcat:qk6nkmsoqbaxjmx5gzxihped3m

Achieving Fairness in Stochastic Multi-armed Bandit Problem [article]

Vishakha Patil, Ganesh Ghalme, Vineet Nair, Y. Narahari
2019 arXiv   pre-print
We investigate the interplay between learning and fairness in terms of a pre-specified vector denoting the fractions of guaranteed pulls.  ...  In particular, when the learning algorithm is UCB1, we show that our algorithm achieves O(log(T)) r-Regret. Finally, we evaluate the cost of fairness in terms of the conventional notion of regret.  ...  Outline of the Paper: In the next section we discuss the related work in the area of fairness in machine learning and fairness in multi-armed bandits in specific.  ... 
arXiv:1905.11260v3 fatcat:7nnynzq24rdxfb66voso5sa7re

Balancing Between Accuracy and Fairness for Interactive Recommendation with Reinforcement Learning [chapter]

Weiwen Liu, Feng Liu, Ruiming Tang, Ben Liao, Guangyong Chen, Pheng Ann Heng
2020 Lecture Notes in Computer Science  
To resolve this problem, we propose a reinforcement learning based framework, FairRec, to dynamically maintain a longterm balance between accuracy and fairness in IRS.  ...  In Interactive Recommender Systems (IRS), user preferences and the system's fairness status are constantly changing over time.  ...  LinUCB [13] is the state-ofthe-art contextual bandits algorithm that sequentially selects items and balances between exploitation and exploration in IRS; (v) DRR.  ... 
doi:10.1007/978-3-030-47426-3_13 fatcat:slo25jndcvb2hj4xbacdsgucmy

Individual Fairness in Hindsight [article]

Swati Gupta, Vijay Kamble
2019 arXiv   pre-print
We introduce two definitions: (i) fairness-across-time (FT) and (ii) fairness-in-hindsight (FH).  ...  We show that these two definitions can have drastically different implications in the setting where the principal needs to learn the utility model.  ...  Celis et al. (2018) consider a contextual bandit problem arising in personalization and address the problem of ensuring another notion of fairness called group fairness 4 across time.  ... 
arXiv:1812.04069v3 fatcat:4ls6vqjvujhvdnl22x36ttryia

Efficient Resource Allocation with Fairness Constraints in Restless Multi-Armed Bandits [article]

Dexun Li, Pradeep Varakantham
2022 arXiv   pre-print
To achieve this goal, we formally define the fairness constraints in RMAB and provide planning and learning methods to solve RMAB in a fair manner.  ...  Restless Multi-Armed Bandits (RMAB) is an apt model to represent decision-making problems in public health interventions (e.g., tuberculosis, maternal, and child care), anti-poaching planning, sensor monitoring  ...  Matthew Joseph, Michael Kearns, Jamie Morgenstern, and Aaron Roth. Fairness in learning: Classic and contextual bandits. arXiv preprint arXiv:1605.07139, 2016.  ... 
arXiv:2206.03883v2 fatcat:oq27vxg4ynfynftwizakbyucga

Individually Fair Learning with One-Sided Feedback [article]

Yahav Bechavod, Aaron Roth
2022 arXiv   pre-print
We then construct an efficient reduction from our problem of online learning with one-sided feedback and a panel reporting fairness violations to the contextual combinatorial semi-bandit problem (Cesa-Bianchi  ...  Finally, we show how to leverage the guarantees of two algorithms in the contextual combinatorial semi-bandit setting: Exp2 (Bubeck et al., 2012) and the oracle-efficient Context-Semi-Bandit-FTPL (Syrgkanis  ...  AR is supported in part by NSF grant FAI-2147212 and the Simons Collaboration on the Theory of Algorithmic Fairness.  ... 
arXiv:2206.04475v1 fatcat:ufxk64hdlbhefdwt4svftnwufu

Fair Server Selection in Edge Computing with Q-Value-Normalized Action-Suppressed Quadruple Q-Learning

Alaa Eddin Alchalabi, Shervin Shirmohammadi, Shady Mohammed, Sorin Stoian, Karthigesu Vijayasuganthan
2021 IEEE Transactions on Artificial Intelligence  
We also introduce action suppression, Quadruple Q-Learning (QQL), and Q-value normalization in RL.  ...  Due to the dynamic and rapidly evolving nature of such an environment and the capacity limitation of the servers, we propose as solution a Reinforcement Learning (RL) method in the form of a Quadruple  ...  The AEN outputs a linear contextual bandit model that eliminates actions with high probability.  ... 
doi:10.1109/tai.2021.3105087 fatcat:tfvqdm6pfzahxltyn63xw522hy

Fairness and Welfare Quantification for Regret in Multi-Armed Bandits [article]

Siddharth Barman, Arindam Khan, Arnab Maiti, Ayush Sawarni
2022 arXiv   pre-print
Focussing on the classic multi-armed bandit (MAB) framework, the current work quantifies the performance of bandit algorithms by applying a fundamental welfare function, namely the Nash social welfare  ...  mean (among the arms) and the algorithm's performance.  ...  Notation and Preliminaries We study the classic (stochastic) multi-armed bandit problem.  ... 
arXiv:2205.13930v1 fatcat:jtoi7qqmwrd3pbpxiimxzmvuaq
« Previous Showing results 1 — 15 out of 1,475 results