Home

Welcome to the Text2ALM wiki!

The goal of this thesis is to design a Narrative Understanding system, Text2ALM, using the methodology that utilizes an action language ALM to allow inferences based on complex interactions of events described in texts. The methodology was originally outlined by Michael Gelfond, Daniela Inclezan, and Yuliya Lierler (2017). This methodology assumes the extension of the VerbNet lexicon with interpretable semantic annotations in ALM and uses the Text2DRS resource developed by Gang Ling (2018). These resources are used to produce ALM system descriptions for input discourses. The Text2DRS system takes a narrative text as input and outputs a discourse representation structure in Neo-Davidsonian style to annotate relevant entities, events, and their relationships. An ALM system description composed with this information will be created and used with a basic commonsense library of ALM modules to generate a formal structure capturing the narrative’s properties.

Publications Related to this Work

• Processing Narratives by Means of Action Languages. 2019 Thesis Report
• Information Extraction Tool Text2ALM: From Narratives to Action Language System Descriptions. ICLP 2019 Report
• An Architecture of Semantic Information Extraction Tool Text2ALM. OSTIS 2020 Report