Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

AI-Generated Compromises for Coalition Formation: Modeling, Simulation, and a Textual Case Study

Eyal Briman, Ehud Shapiro, Nimrod Talmon

TL;DR

The paper addresses the problem of forming majority-supported compromises among agents by extending coalition-formation into a flexible, bounded-rationality framework. It introduces mediators that generate compromise proposals within a semantic metric space, enabling AI-assisted, democratic text editing such as constitution drafting. Concrete realizations in both Euclidean and textual domains are developed, leveraging LLMs (GPT-3.5) and the Universal Sentence Encoder to embed and optimize text proposals. Through extensive simulations, the authors demonstrate convergence to majority-supported documents and reveal how mediator choice, agent rationality, and space (Euclidean vs. textual) impact convergence speed and draft quality. Overall, the work shows AI-mediated mediation can enhance collaborative drafting tasks at scale, with implications for participatory governance and large-scale democratic editing.

Abstract

The challenge of finding compromises between agent proposals is fundamental to AI sub-fields such as argumentation, mediation, and negotiation. Building on this tradition, Elkind et al. (2021) introduced a process for coalition formation that seeks majority-supported proposals preferable to the status quo, using a metric space where each agent has an ideal point. The crucial step in this iterative process involves identifying compromise proposals around which agent coalitions can unite. How to effectively find such compromise proposals, however, remains an open question. We address this gap by formalizing a holistic model that encompasses agent bounded rationality and uncertainty and developing AI models to generate such compromise proposals. We focus on the domain of collaboratively writing text documents -- e.g., to enable the democratic creation of a community constitution. We apply NLP (Natural Language Processing) techniques and utilize LLMs (Large Language Models) to create a semantic metric space for text and develop algorithms to suggest suitable compromise points. To evaluate the effectiveness of our algorithms, we simulate various coalition formation processes and demonstrate the potential of AI to facilitate large-scale democratic text editing, such as collaboratively drafting a constitution, an area where traditional tools are limited.

AI-Generated Compromises for Coalition Formation: Modeling, Simulation, and a Textual Case Study

TL;DR

The paper addresses the problem of forming majority-supported compromises among agents by extending coalition-formation into a flexible, bounded-rationality framework. It introduces mediators that generate compromise proposals within a semantic metric space, enabling AI-assisted, democratic text editing such as constitution drafting. Concrete realizations in both Euclidean and textual domains are developed, leveraging LLMs (GPT-3.5) and the Universal Sentence Encoder to embed and optimize text proposals. Through extensive simulations, the authors demonstrate convergence to majority-supported documents and reveal how mediator choice, agent rationality, and space (Euclidean vs. textual) impact convergence speed and draft quality. Overall, the work shows AI-mediated mediation can enhance collaborative drafting tasks at scale, with implications for participatory governance and large-scale democratic editing.

Abstract

The challenge of finding compromises between agent proposals is fundamental to AI sub-fields such as argumentation, mediation, and negotiation. Building on this tradition, Elkind et al. (2021) introduced a process for coalition formation that seeks majority-supported proposals preferable to the status quo, using a metric space where each agent has an ideal point. The crucial step in this iterative process involves identifying compromise proposals around which agent coalitions can unite. How to effectively find such compromise proposals, however, remains an open question. We address this gap by formalizing a holistic model that encompasses agent bounded rationality and uncertainty and developing AI models to generate such compromise proposals. We focus on the domain of collaboratively writing text documents -- e.g., to enable the democratic creation of a community constitution. We apply NLP (Natural Language Processing) techniques and utilize LLMs (Large Language Models) to create a semantic metric space for text and develop algorithms to suggest suitable compromise points. To evaluate the effectiveness of our algorithms, we simulate various coalition formation processes and demonstrate the potential of AI to facilitate large-scale democratic text editing, such as collaboratively drafting a constitution, an area where traditional tools are limited.

Paper Structure

This paper contains 29 sections, 3 equations, 1 figure, 1 table.

Table of Contents

  1. Introduction
  2. Generalizing Agent Behavior.
  3. Collaborative Text Editing and the Mediator Concept.
  4. Main Contributions.
  5. Model State
  6. Initialization, Iterative Process, and Halting Conditions
  7. Initialization
  8. Process
  9. A halting condition
  10. Concrete Model Realizations
  11. One Concrete Agent Model
  12. Two Concrete Constitutions
  13. Several Concrete Mediators
  14. Choosing the coalitions $d_i, d_j$
  15. Choosing the Compromise Point $p$
  16. ...and 14 more sections

Figures (1)

  • Figure 1: Coalition formation result- "Dealing with Global Warming", for $n=10, C=\text{False}, \alpha=0, \sigma=0, I=\text{True}$.

Theorems & Definitions (13)

  • Definition 1: mediator
  • Definition 2: Agent, vote
  • Definition 3: Constitution
  • Definition 4: A deterministic agent model
  • Definition 5: A probabilistic agent model
  • Remark 6
  • Remark 7
  • Remark 8
  • Remark 9
  • Remark 10
  • ...and 3 more