AAAI Publications, Sixteenth International Conference on Principles of Knowledge Representation and Reasoning

Font Size: 
Bayesian Model Selection in Statistical Construction of Justification
Hiroyuki Kido

Last modified: 2018-09-24

Abstract


Argumentation mining involves identification of an attack relation between natural language sentences. Bayesian inference characterizing argument-based reasoning addresses this issue by calculating the posterior distribution over attack relations given acceptability statuses of arguments. This paper discusses the use of Bayesian model selection where graph-theoretic properties impose restrictions on the graphic structure of attack relations.

Keywords


argumentation frameworks; argumentation mining; empirical Bayes

Full Text: PDF