Daily Arxiv

This is a page that curates AI-related papers published worldwide.
All content here is summarized using Google Gemini and operated on a non-profit basis.
Copyright for each paper belongs to the authors and their institutions; please make sure to credit the source when sharing.

AI-Generated Compromises for Coalition Formation

Created by
  • Haebom

Author

Eyal Briman, Ehud Shapiro, Nimrod Talmon

Outline

This paper addresses the problem of finding compromises between agent proposals. Specifically, we propose a model that considers agents' bounded rationality and uncertainty, building on the coalition formation process of Elkind et al. (2021) to find majority-supported proposals. Focusing on collaborative document creation, such as drafting a community constitution, we use natural language processing techniques and large-scale language models to derive a semantic metric space within the text and design an algorithm that proposes compromises likely to receive widespread support. Simulations demonstrate that AI can enable large-scale democratic text editing.

Takeaways, Limitations

•
Takeaways:
◦
Contributes to solving realistic problems by presenting a compromise generation model that takes into account the agent's limited rationality and uncertainty.
◦
Leveraging natural language processing and large-scale language models to increase applicability to complex situations such as real-world document writing.
◦
Providing AI-based solutions to challenging problems such as large-scale democratic text editing.
•
Limitations:
◦
Further experiments and verification are needed to verify the generalizability of the proposed model and algorithm.
◦
Focused on a specific domain (collaborative document creation), with potential for expansion to other fields.
◦
There is a need to analyze the impact of biases and limitations of the large-scale language models used on the results.
👍
Made with Slashpage