ASU at TextGraphs 2019 Shared Task: Explanation ReGeneration using Language Models and Iterative Re-Ranking

TL;DR

This paper presents a system for Explanation Regeneration using language models and iterative re-ranking, achieving second place in the TextGraphs 2019 Shared Task with a MAP of 41.3%.

Contribution

The work introduces a novel ranking-based approach employing language models and iterative re-ranking for explanation regeneration tasks.

Findings

Achieved 2nd place in the shared task

Attained a mean average precision of 41.3%

Demonstrated effectiveness of iterative re-ranking

Abstract

In this work we describe the system from Natural Language Processing group at Arizona State University for the TextGraphs 2019 Shared Task. The task focuses on Explanation Regeneration, an intermediate step towards general multi-hop inference on large graphs. Our approach consists of modeling the explanation regeneration task as a \textit{learning to rank} problem, for which we use state-of-the-art language models and explore dataset preparation techniques. We utilize an iterative re-ranking based approach to further improve the rankings. Our system secured 2nd rank in the task with a mean average precision (MAP) of 41.3\% on the test set.