AdaLoGN: Adaptive Logic Graph Network for Reasoning-Based Machine Reading Comprehension

TL;DR

AdaLoGN is a neural-symbolic model that adaptively constructs and extends logical relation graphs for improved reasoning in machine reading comprehension, demonstrating promising results on challenging datasets.

Contribution

The paper introduces AdaLoGN, a novel adaptive logic graph network that combines neural and symbolic reasoning with a subgraph-to-node message passing mechanism.

Findings

Achieves promising results on ReClor and LogiQA datasets.

Effectively infers and extends logical relations in text.

Enhances context-option interaction for multiple-choice questions.

Abstract

Recent machine reading comprehension datasets such as ReClor and LogiQA require performing logical reasoning over text. Conventional neural models are insufficient for logical reasoning, while symbolic reasoners cannot directly apply to text. To meet the challenge, we present a neural-symbolic approach which, to predict an answer, passes messages over a graph representing logical relations between text units. It incorporates an adaptive logic graph network (AdaLoGN) which adaptively infers logical relations to extend the graph and, essentially, realizes mutual and iterative reinforcement between neural and symbolic reasoning. We also implement a novel subgraph-to-node message passing mechanism to enhance context-option interaction for answering multiple-choice questions. Our approach shows promising results on ReClor and LogiQA.