{"title":"通过基础模型自我解决的自然语言机制","authors":"Nicolas Della Penna","doi":"arxiv-2407.07845","DOIUrl":null,"url":null,"abstract":"Practical mechanisms often limit agent reports to constrained formats like\ntrades or orderings, potentially limiting the information agents can express.\nWe propose a novel class of mechanisms that elicit agent reports in natural\nlanguage and leverage the world-modeling capabilities of large language models\n(LLMs) to select outcomes and assign payoffs. We identify sufficient conditions\nfor these mechanisms to be incentive-compatible and efficient as the LLM being\na good enough world model and a strong inter-agent information\nover-determination condition. We show situations where these LM-based\nmechanisms can successfully aggregate information in signal structures on which\nprediction markets fail.","PeriodicalId":501316,"journal":{"name":"arXiv - CS - Computer Science and Game Theory","volume":"168 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Natural Language Mechanisms via Self-Resolution with Foundation Models\",\"authors\":\"Nicolas Della Penna\",\"doi\":\"arxiv-2407.07845\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Practical mechanisms often limit agent reports to constrained formats like\\ntrades or orderings, potentially limiting the information agents can express.\\nWe propose a novel class of mechanisms that elicit agent reports in natural\\nlanguage and leverage the world-modeling capabilities of large language models\\n(LLMs) to select outcomes and assign payoffs. We identify sufficient conditions\\nfor these mechanisms to be incentive-compatible and efficient as the LLM being\\na good enough world model and a strong inter-agent information\\nover-determination condition. We show situations where these LM-based\\nmechanisms can successfully aggregate information in signal structures on which\\nprediction markets fail.\",\"PeriodicalId\":501316,\"journal\":{\"name\":\"arXiv - CS - Computer Science and Game Theory\",\"volume\":\"168 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-07-10\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - CS - Computer Science and Game Theory\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2407.07845\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - CS - Computer Science and Game Theory","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2407.07845","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Natural Language Mechanisms via Self-Resolution with Foundation Models
Practical mechanisms often limit agent reports to constrained formats like
trades or orderings, potentially limiting the information agents can express.
We propose a novel class of mechanisms that elicit agent reports in natural
language and leverage the world-modeling capabilities of large language models
(LLMs) to select outcomes and assign payoffs. We identify sufficient conditions
for these mechanisms to be incentive-compatible and efficient as the LLM being
a good enough world model and a strong inter-agent information
over-determination condition. We show situations where these LM-based
mechanisms can successfully aggregate information in signal structures on which
prediction markets fail.