A Probabilistic Model with Commonsense Constraints for Pattern-based Temporal Fact Extraction

TL;DR

This paper introduces a probabilistic graphical model that improves temporal fact extraction from unstructured text by automatically inferring fact accuracy and pattern reliability without supervision, leveraging temporal signals and commonsense constraints.

Contribution

It presents a novel generative probabilistic model that incorporates temporal signals and commonsense constraints for more accurate temporal fact extraction without supervision.

Findings

Significantly outperforms existing methods on news data

Effectively models pattern reliability using temporal signals

Automatically infers true facts and pattern reliability

Abstract

Textual patterns (e.g., Country's president Person) are specified and/or generated for extracting factual information from unstructured data. Pattern-based information extraction methods have been recognized for their efficiency and transferability. However, not every pattern is reliable: A major challenge is to derive the most complete and accurate facts from diverse and sometimes conflicting extractions. In this work, we propose a probabilistic graphical model which formulates fact extraction in a generative process. It automatically infers true facts and pattern reliability without any supervision. It has two novel designs specially for temporal facts: (1) it models pattern reliability on two types of time signals, including temporal tag in text and text generation time; (2) it models commonsense constraints as observable variables. Experimental results demonstrate that our model significantly outperforms existing methods on extracting true temporal facts from news data.