RI: Small: Protecting Social Choice Mechanisms from Malicious Influence

社会选择机制对恶意影响的脆弱性是一个长期关注的问题。通过使用虚假或极具误导性的信息，数字媒体平台成为一种日益重要的恶意影响手段。这种影响的净效果是改变机制参与者对问题相对重要性的态度，以及他们对候选人和他们的观点的看法，从而改变社会选择结果。这项提议旨在开发系统的、定量的和算法的方法，用于研究操纵以对问题的看法和观点为目标时对社会选择机制的控制。因此，这项研究有可能对更广泛的社会选择机制对恶意影响的脆弱性的跨学科研究做出重大贡献。具体地说，本研究将探索一种控制社会选择机制的新模型，该模型建立在投票的空间理论基础上，在该理论中，选民对候选人的偏好是基于他们在问题上的相对相似性而产生的。在这个模型中，操纵的主要目标是对问题的感知，无论是从它们的相对重要性来看，还是对候选人观点的感知。在算法上，这打开了几个以前几乎没有研究的新问题类别，例如问题选择问题，恶意一方可能会选择哪些问题最突出，或者意见操纵，其中恶意行为者旨在就候选人在问题上的立场误导机制参与者。从概念上讲，这个问题与现实中常见的颠覆社会选择机制的手段有更直接的联系，例如在社交媒体和传统媒体上的广告和竞选活动。此外，该项目将研究空间模型中保护社会选择机制的双重问题，为这一问题提供一系列新颖的建模和算法贡献。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Vulnerability of social choice mechanisms to malicious influence is a long-standing concern. An increasingly important means of malicious influence involves digital media platforms through the use of fake or highly misleading information. The net effect of such influence is to change the attitudes of mechanism participants about the relative importance of issues, as well as their perception of candidates and their views and, as a consequence, change social choice outcomes. This proposal aims to develop systematic quantitative and algorithmic approaches for studying control of social choice mechanisms when manipulation targets perceptions about and views on issues. Consequently, this research has the potential to significantly contribute to the broader interdisciplinary study of vulnerability of social choice mechanisms to malicious influence. Specifically, this research will investigate a novel model of control of social choice mechanisms which builds on the spatial theory of voting in which voters' preferences over candidates are generated based on their relative similarity on issues. In this model, the primary target of manipulation is perception of issues, both in terms of their relative importance, as well as perception of candidates' views. Algorithmically, this opens several new classes of problems with little prior research, such as the problem of issue selection, where the malicious party may select which issues are most salient, or opinion manipulation where a malicious actor aims to mislead mechanism participants about where the candidates stand on issues. Conceptually, this problem has a more direct connection to common means of subversion of social choice mechanism in reality, such as advertising and campaigning, both on social and conventional media. Furthermore, this project will study the dual problem of protecting social choice mechanisms in the spatial model, providing a series of novel modeling and algorithmic contributions for this problem.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.