More Like this

1. Balancing Competing Objectives with Noisy Data: Score-Based Classifiers for Welfare-Aware Machine Learning

   Rolf, Esther; Simchowitz, Max; Dean, Sarah; Liu, Lydia; Bjorkegren, Daniel; Hardt, Moritz; Blumenstock, Joshua E (January 2020, Proceedings of Machine Learning Research)

   null (Ed.)

   While real-world decisions involve many competing objectives, algorithmic decisions are often evaluated with a single objective function. In this paper, we study algorithmic policies which explicitly trade off between a private objective (such as profit) and a public objective (such as social welfare). We analyze a natural class of policies which trace an empirical Pareto frontier based on learned scores, and focus on how such decisions can be made in noisy or data-limited regimes. Our theoretical results characterize the optimal strategies in this class, bound the Pareto errors due to inaccuracies in the scores, and show an equivalence between optimal strategies and a rich class of fairness-constrained profit-maximizing policies. We then present empirical results in two different contexts — online content recommendation and sustainable abalone fisheries — to underscore the generality of our approach to a wide range of practical decisions. Taken together, these results shed light on inherent trade-offs in using machine learning for decisions that impact social welfare. 

   more » « less
2. Disaggregated Interventions to Reduce Inequality

   https://doi.org/10.1145/3465416.3483286  

   Bynum, Lucius; Loftus, Joshua; Stoyanovich, Julia (October 2021, Equity and Access in Algorithms, Mechanisms, and Optimization (EAAMO ’21))

   A significant body of research in the data sciences considers unfair discrimination against social categories such as race or gender that could occur or be amplified as a result of algorithmic decisions. Simultaneously, real-world disparities continue to exist, even before algorithmic decisions are made. In this work, we draw on insights from the social sciences brought into the realm of causal modeling and constrained optimization, and develop a novel algorithmic framework for tackling pre-existing real-world disparities. The purpose of our framework, which we call the “impact remediation framework,” is to measure real-world disparities and discover the optimal intervention policies that could help improve equity or access to opportunity for those who are underserved with respect to an outcome of interest. We develop a disaggregated approach to tackling pre-existing disparities that relaxes the typical set of assumptions required for the use of social categories in structural causal models. Our approach flexibly incorporates counterfactuals and is compatible with various ontological assumptions about the nature of social categories. We demonstrate impact remediation with a hypothetical case study and compare our disaggregated approach to an existing state-of-the-art approach, comparing its structure and resulting policy recommendations. In contrast to most work on optimal policy learning, we explore disparity reduction itself as an objective, explicitly focusing the power of algorithms on reducing inequality. 

   more » « less
3. Pricing in Prosumer Aggregations using Reinforcement Learning

   https://doi.org/10.1145/3447555.3464853  

   Agwan, Utkarsha; Spangher, Lucas; Arnold, William; Srivastava, Tarang; Poolla, Kameshwar; Spanos, Costas J. (June 2021, e-Energy '21: Proceedings of the Twelfth ACM International Conference on Future Energy Systems)

   Prosumers with generation and storage capabilities can supply en- ergy back to the grid, or trade their surplus with other prosumers for their mutual benefit. A prosumer aggregation that facilitates such trades will price the energy being traded to achieve an objective such as profit maximization, social welfare, or market equilibrium. We propose the use of reinforcement learning to design a trans- active controller to price energy in a prosumer aggregation. This has an advantage over other decentralized pricing mechanisms as it does not rely on iterative price settlement or load estimation by prosumers, and estimates the price in a day ahead manner. We present numerical case studies to evaluate our controller, and dis- cuss extensions to implement this in real prosumer aggregations. 

   more » « less
4. "I'm not sure what difference is between their content and mine, other than the person itself": A Study of Fairness Perception of Content Moderation on YouTube

   https://doi.org/10.1145/3555150  

   Ma, Renkai; Kou, Yubo (November 2022, Proceedings of the ACM on Human-Computer Interaction)

   How social media platforms could fairly conduct content moderation is gaining attention from society at large. Researchers from HCI and CSCW have investigated whether certain factors could affect how users perceive moderation decisions as fair or unfair. However, little attention has been paid to unpacking or elaborating on the formation processes of users' perceived (un)fairness from their moderation experiences, especially users who monetize their content. By interviewing 21 for-profit YouTubers (i.e., video content creators), we found three primary ways through which participants assess moderation fairness, including equality across their peers, consistency across moderation decisions and policies, and their voice in algorithmic visibility decision-making processes. Building upon the findings, we discuss how our participants' fairness perceptions demonstrate a multi-dimensional notion of moderation fairness and how YouTube implements an algorithmic assemblage to moderate YouTubers. We derive translatable design considerations for a fairer moderation system on platforms affording creator monetization. 

   more » « less
5. EFx Budget-Feasible Allocations with High Nash Welfare

   Garbea, Marius; Gkatzelis, Vasilis; Tan, Xizhi (October 2023, 26th European Conference on Artificial Intelligence)

   We study the problem of allocating indivisible items to budget-constrained agents, aiming to provide fairness and efficiency guarantees. Specifically, our goal is to ensure that the resulting allocation is envy-free up to any item (EFx) while minimizing the amount of inefficiency that this needs to introduce. We first show that there exist two-agent problem instances for which no EFx allocation is Pareto-efficient. We, therefore, turn to approximation and use the (Pareto-efficient) maximum Nash welfare allocation as a benchmark. For two-agent instances, we provide a procedure that always returns an EFx allocation while achieving the best possible approximation of the optimal Nash social welfare that EFx allocations can achieve. For the more complicated case of three-agent instances, we provide a procedure that guarantees EFx, while achieving a constant approximation of the optimal Nash social welfare for any number of items. 

   more » « less