Representative Social Choice: From Learning Theory to AI Alignment
Keywords: Language Model Alignment, Social Choice Theory, Statistical Learning Theory
TL;DR: Using tools from ML theory, we introduce the representative social choice framework for the modeling of democratic representation in collective decisions, with key applications in the alignment of language models.
Abstract: Social choice theory is the study of preference aggregation across a population, used both in mechanism design for human agents and in the democratic alignment of language models.
In this study, we propose the *representative social choice* framework for the modeling of democratic representation in collective decisions, where the number of issues and individuals are too large for mechanisms to consider all preferences directly. These scenarios are widespread in real-world decision-making processes, such as jury trials, indirect elections, legislation processes, corporate governance, and, more recently, language model alignment.
In representative social choice, the population is *represented* by a finite sample of individual-issue pairs, based on which social choice decisions are made. We show that many of the deepest questions in representative social choice can be naturally formulated as statistical learning problems, and prove the generalization properties of social choice mechanisms using the theory of statistical machine learning. We further formulate axioms for representative social choice, and prove Arrow-like impossibility theorems with new combinatorial tools of analysis. Our framework introduces the representative approach to social choice, and opens up research directions at the intersection of social choice, learning theory, and AI alignment.
Submission Number: 62
Loading