CRII: III: Novel Computational Social Choice Extensions for Highly Distributed Decision-Making Contexts

点击翻译按钮获取中文摘要

Over the last two decades, there has been a growing interest in the aggregation of individual preferences into socially desirable collective choices (e.g., crowdsourced recommendations, online voting), helping to propel the new interdisciplinary field of computational social choice. In many ways, the emphasis on examining whether and how preference aggregation algorithms can be designed to ensure fairness, avoid strategic manipulation, and achieve other socially desirable properties is driven by the largely unchecked prevalence of automated "black-box" decision-making technologies within everyday life. While the rising interest in this new field has resulted in various landmark results, implementation of the more socially beneficial methodologies within modern contexts remains severely limited due to a combination of incompatible assumptions and computational difficulties. This research project will seek to extend the real-world applicability of these robust methodologies by melding socio-theoretical insights, efficient algorithms, and advanced operations research techniques. Accordingly, this novel approach will build interdisciplinary bridges with computer science and expose computational social choice to new audiences. Moreover, through an overarching emphasis on rigorous theoretical underpinnings, the envisioned contributions will address the pressing need to develop and implement interpretable decision-making algorithms. Hence, the outcomes of this project will prospectively have widespread impacts on society. The advances envisioned through the completion of this project will expand the traditional scope of computational social choice, particularly of Kemeny aggregation, which is widely regarded as one of the most robust preference-ranking aggregation frameworks in the literature. The focus of this research project will be on highly distributed decision-making contexts, which are often characterized by large numbers of alternatives, tied (i.e., partial) preferences, errors, and/or incompleteness. This will be accomplished by exploring symbiotic relationships between social choice theory, efficient algorithms, and operations research techniques. Planned research tasks will include: (i) Establishing social choice axioms and properties that different distance measures should satisfy when dealing with partial and incomplete preference rankings; (ii) Constructing mathematical models and decomposition algorithms that take advantage of these insights; and (iii) Exploring the validity and pragmatic implications of these measures via formal statistical methods and benchmark instances of preference data.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.