Social Choice Theory

By Amartya Sen. Self-Organising Multi-Agent Systems, 2021

380 citations recorded by the source corpus.

Social choice theory extends information economics to the context of group or societal decision-making. The underlying assumption is that we do not know the preferences of individuals, but we want to use them to make decisions. The overarching question is whether it is possible to design decision-making rules or systems that will both encourage individuals to reveal their preferences and then use that information to make a reasonable decision. The general approach to the problem is …rst to study decision rules in the case of complete information about preferences and to identify a set of rules that satis…es particular desirable properties. Then, we turn back to the question of incentives for individuals to reveal information about their preferences to determine when it is a Nash equilibrium to reveal information honestly in voting (as an example of a rule for how societies make decisions in practice). An important observation is that the incentives to reveal information are bound up completely with how that information will be used in the societal decision. In general, a societal decision rule that has good properties for the case of complete information will also provide incentives for individuals to reveal their preferences honestly in voting. However, a societal decision rule that does not have good properties for the case of complete information about individual preferences, will provide incentives for people to conceal their preferences, i.e. by voting "dishonestly". Thus it makes sense to start with the decision rule and work backwards to incentives. There is another underlying assumption, which is that the attributes of available alternatives are known to everyone. It is not possible, for example, that one person is an expert whose opinion is more informed than everyone else's. The only uncertainty is how individuals value the alternatives. Throughout this discussion, we will (mostly) sidestep the question of strength of preferences. We 175

Subfields: General Decision Sciences

View Paper: https://www.semanticscholar.org/paper/ce11759cc4047695a847c7def648e5ad9ef0dbf7

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In General Decision Sciences

1. Didier Dubois
2. Henri Prade
3. F. Delgado
4. Stephen Yang
5. Qian Yang
6. Michael Madaio

Citations in the corpus

• Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback (2023, positive): “As such, it is important to analyze RLHF from the lens of social choice theory [Sen, 1986] and work to understand whether the means by which it aggregates preferences is normatively acceptable.”
• Managing extreme AI risks amid rapid progress (2023, methodology): “Inclusive AI development: AI advancement will need methods to mitigate biases and integrate the values of the many populations it will affect 20,55 .”
• On the use of aggregation operations in information fusion processes (2004, background): “In multiagent fusion the tradition of social choice is to adopt purely ordinal approaches, leading to impossibility theorems (like Arrow’s), that scholars try to bypass by Fnding suitably relaxed settings [57].”
• The Participatory Turn in AI Design: Theoretical Foundations and the Current State of Practice (2023, methodology): “SCT focuses on identifying people’s preferences for (e.g.) public policies and mathematically aggregating those in a preference-ranking model [e.g., 13, 108, 184, 195, 196].”