Filters








117,299 Hits in 6.1 sec

Algorithmic decision making and the cost of fairness [article]

Sam Corbett-Davies, Emma Pierson, Avi Feller, Sharad Goel, Aziz Huq
2017 arXiv   pre-print
Because the optimal constrained and unconstrained algorithms generally differ, there is tension between improving public safety and satisfying prevailing notions of algorithmic fairness.  ...  We focus on algorithms for pretrial release decisions, but the principles we discuss apply to other domains, and also to human decision makers carrying out structured decision rules.  ...  Knight Foundation, and by the Hellman Fellows Fund.  ... 
arXiv:1701.08230v3 fatcat:5jewivcsm5azbgplrgbodsyzc4

Algorithmic Decision Making and the Cost of Fairness

Sam Corbett-Davies, Emma Pierson, Avi Feller, Sharad Goel, Aziz Huq
2017 Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining - KDD '17  
Because the optimal constrained and unconstrained algorithms generally dier, there is tension between improving public safety and satisfying prevailing notions of algorithmic fairness.  ...  We focus on algorithms for pretrial release decisions, but the principles we discuss apply to other domains, and also to human decision makers carrying out structured decision rules.  ...  Knight Foundation, and by the Hellman Fellows Fund. Data and code to reproduce our results are available at hps://github.com/5harad/cost-of-fairness.  ... 
doi:10.1145/3097983.3098095 dblp:conf/kdd/Corbett-DaviesP17 fatcat:qdk7cofj7fdojnbhf5k2mmjln4

A Harm-Reduction Framework for Algorithmic Fairness

Micah Altman, Alexandra Wood, Effy Vayena
2018 IEEE Security and Privacy  
In this article we recognize the profound effects that algorithmic decision-making can have on people's lives and propose a harm-reduction framework for algorithmic fairness.  ...  Also, an algorithmic decision is unfair when it is regressive, i.e., when members of disadvantaged groups pay a higher cost for the social benefits of that decision.  ...  A nominally fair algorithm will always yield an even distribution of costs across groups when each of the probabilities, percentages, and costs above are equal for each group.  ... 
doi:10.1109/msp.2018.2701149 fatcat:y4yjt4z4wff2vjsnwonijzskhq

Fairness On The Ground: Applying Algorithmic Fairness Approaches to Production Systems [article]

Chloé Bakalar, Renata Barreto, Stevie Bergman, Miranda Bogen, Bobbie Chern, Sam Corbett-Davies, Melissa Hall, Isabel Kloumann, Michelle Lam, Joaquin Quiñonero Candela, Manish Raghavan, Joshua Simons (+4 others)
2021 arXiv   pre-print
This paper presents an example of one team's approach to the challenge of applying algorithmic fairness approaches to complex production systems within the context of a large technology company.  ...  Many technical approaches have been proposed for ensuring that decisions made by machine learning systems are fair, but few of these proposals have been stress-tested in real-world systems.  ...  Finally, we discussed that approach in two archetypal binary decision-making contexts: algorithmic decision making and human labeling.  ... 
arXiv:2103.06172v2 fatcat:e6lrzvjetfgxjkq7ucyixxdoei

Fair and Unbiased Algorithmic Decision Making: Current State and Future Challenges [article]

Songül Tolan
2019 arXiv   pre-print
In the future, research in algorithmic decision making systems should be aware of data and developer biases and add a focus on transparency to facilitate regular fairness audits.  ...  This has led to a call for fairness-aware machine learning. However, fairness is a complex concept which is also reflected in the attempts to formalize fairness for algorithmic decision making.  ...  Context We review the problem of discrimination and bias in algorithmic decision making.  ... 
arXiv:1901.04730v1 fatcat:exsvq4g52vg3rlyfomp3tszy6e

Managing large dynamic graphs efficiently

Jayanta Mondal, Amol Deshpande
2012 Proceedings of the 2012 international conference on Management of Data - SIGMOD '12  
Second, we propose a clustering-based approach to amortize the costs of making these replication decisions.  ...  Finally, we propose using a fairness criterion to dictate how replication decisions should be made.  ...  Acknowledgments: This work was supported by Air Force Research Lab (AFRL) under contract FA8750-10-C-0191, by NSF under grant IIS-0916736, and an Amazon AWS in Education Research grant.  ... 
doi:10.1145/2213836.2213854 dblp:conf/sigmod/MondalD12 fatcat:mzeru22b6ff3pj34oefuehzpa4

Modeling participation behavior in repeated task allocations with fuzzy connectives

Qing Chuan Ye, Yingqian Zhang, Uzay Kaymak
2017 2017 IEEE International Conference on Systems, Man, and Cybernetics (SMC)  
When making use of the fair allocation algorithm, we can see that the decline in the number of participants is not nearly as sharp as with the minimum-cost algorithm.  ...  We already observe a sharp decline in the number of participants in the second round when making use of the minimum-cost algorithm, with both homogeneous and heterogeneous costs.  ... 
doi:10.1109/smc.2017.8123124 dblp:conf/smc/YeZK17 fatcat:3x37c4riargrngo2h233jdecqm

Contrastive Fairness in Machine Learning [article]

Tapabrata Chakraborti, Arijit Patra, Alison Noble
2019 arXiv   pre-print
We introduce concepts and mathematical tools using causal inference to address contrastive fairness in algorithmic decision-making with illustrative examples.  ...  However research in fairness of algorithms has focused on the counterfactual questions "what if?" or "why?"  ...  He is one of the authors of the NeurIPS 2017 paper on Counterfactual Fairness [9] and he kindly shared his data and code with us.  ... 
arXiv:1905.07360v4 fatcat:ovmwfzzsard3fmwjqryi4lwsy4

Provably Fair Representations [article]

Daniel McNamara, Cheng Soon Ong, Robert C. Williamson
2017 arXiv   pre-print
This has led to considerable interest in making such machine learning systems fair. One approach is to transform the input data used by the algorithm.  ...  We formally define the 'cost of mistrust' of using this model compared to the setting where there is a single trusted party, and provide bounds on this cost in particular cases.  ...  input data, a data user who makes decisions from the data, and a data regulator who oversees fair use of the data.  ... 
arXiv:1710.04394v1 fatcat:jfao7a6jdfdytgagyi4umzktzu

Integer Search Algorithm: A New Discrete Multi-Objective Algorithm for Pavement Maintenance Management Optimization

Abdulraaof Alqaili, Mohammed Qais, Abdullah Al-Mansour
2021 Applied Sciences  
The ISA and genetic algorithm (GA) are applied to improve the performance condition rating (PCR) of the pavement in developing countries, where the annual budget is limited, so a minimum cost for three  ...  Optimization techniques keep road performance at a good level using a cost-effective maintenance strategy. Thus, the trade-off between cost and road performance is a multi-objective function.  ...  Conflicts of Interest: The authors declare no conflict of interest.  ... 
doi:10.3390/app11157170 fatcat:6yawgeininfbjcuxz35acx57o4

Contrastive Fairness in Machine Learning

Tapabrata Chakraborti, Arijit Patra, J. Alison Noble
2020 Letters of the IEEE Computer Society  
We introduce concepts and mathematical tools using causal inference to address contrastive fairness in algorithmic decision-making with illustrative examples.  ...  However research in fairness of algorithms has focused on the counterfactual questions "what if?" or "why?"  ...  Given the broad use of machine learning algorithms in the modern world, precautions to ensure the fairness of the decision making process of such algorithms is of great importance.  ... 
doi:10.1109/locs.2020.3007845 fatcat:slmi34ip7nc5hc4ffrq2ifaefe

Discriminated by an algorithm: a systematic review of discrimination and fairness by algorithmic decision-making in the context of HR recruitment and HR development

Alina Köchling, Marius Claus Wehner
2020 Business Research  
While firms implement algorithmic decision-making to save costs as well as increase efficiency and objectivity, algorithmic decision-making might also lead to the unfair treatment of certain groups of  ...  Current knowledge about the threats of unfairness and (implicit) discrimination by algorithmic decision-making is mostly unexplored in the human resource management context.  ...  Acknowledgements We thank Maike Giefers, Hannah Kaiser, and Anna Nieter, and Shirin Riazy for their support. Funding Not applicable for that section.  ... 
doi:10.1007/s40685-020-00134-w fatcat:wdhbnwuecragzpfslr2fymhs7i

Legally grounded fairness objectives [article]

Dylan Holden-Sim and Gavin Leech and Laurence Aitchison
2020 arXiv   pre-print
of the social cost.  ...  Here, we formulate Legally Grounded Fairness Objectives (LGFO), which uses signals from the legal system to non-arbitrarily measure the social cost of a specific degree of unfairness.  ...  It also allows for stakeholders other than the technical team to contribute to the system design, and makes use of long-standing legal expertise on decision-making in complex social situations.  ... 
arXiv:2009.11677v1 fatcat:gkll4reaefaptbh7elz4gcbmzy

QoS Differentiated and Fair Packet Scheduling in Broadband Wireless Access Networks

Rong Yu, Yan Zhang, Shengli Xie
2009 EURASIP Journal on Wireless Communications and Networking  
and longterm fairness.  ...  Specifically, we formulate the packet scheduling problem as an average cost Semi-Markov Decision Process (SMDP). Then, we solve the SMDP by using reinforcement learning.  ...  Acknowledgment The work in this paper is supported by programs of NSFC under Grant nos. 60903170, U0835003, and U0635001.  ... 
doi:10.1155/2009/482764 fatcat:hkbbstxurngulk6f6dsmknasfu

Fair, Transparent, and Accountable Algorithmic Decision-making Processes

Bruno Lepri, Nuria Oliver, Emmanuel Letouzé, Alex Pentland, Patrick Vinck
2017 Philosophy & Technology  
In this paper we provide an overview of available technical solutions to enhance fairness, accountability and transparency in algorithmic decision-making.  ...  decision-making processes designed to maximize fairness and transparency.  ...  They show that learning algorithms can be proven to be fair in such a way that the cost (from the perspective of rate of convergence to an optimal decision) of adding fairness to the algorithm is small  ... 
doi:10.1007/s13347-017-0279-x fatcat:jfafkeyn7bdk3kv3w4qhufogu4
« Previous Showing results 1 — 15 out of 117,299 results