辩论中推荐最具包容性的反对与支持论据
人工智能
2014-11-21 v1
摘要
论据是 DirectDemocracyP2P 中的基本对象，它们既可以与请愿书的签名关联出现，也可以与其他辩论性决策（如按重要性排序的缺陷分类）关联出现。签名者就某一议题的论据被归入一个单一理由陈述，按签名类型（例如支持或反对）分类，并可进行多种形式的线索化组织。给定可用输入，所解决的两个问题是：(i) 如何向新投票者推荐给定类型的最佳理由陈述，(ii) 如何推荐一个紧凑的理由陈述列表，以涵盖支持（或反对）某一议题的多数已知论据。我们研究了基于加权二分图的解决方案。
引用
@article{arxiv.1411.5416,
title = {Recommending the Most Encompassing Opposing and Endorsing Arguments in Debates},
author = {Marius C. Silaghi and Roussi Roussev},
journal= {arXiv preprint arXiv:1411.5416},
year = {2014}
}
备注
10 pages. This report was reviewed by a committee within Florida Tech during April 2014, and had been written in Summer 2013 by summarizing a set of emails exchanged during Spring 2013, concerning the DirectDemocracyP2P.net system