Reasoning-Oriented and Analogy-Based Methods for Locating and Editing in Zero-Shot Event-Relational Reasoning

TL;DR

This paper introduces ROLE and ABLE, two methods that improve zero-shot event-relational reasoning by enhancing interpretability, reducing computational costs, and exploiting task similarities, leading to state-of-the-art results.

Contribution

The paper presents novel reasoning-oriented and analogy-based locating and editing methods that optimize zero-shot reasoning in NLP tasks.

Findings

ROLE enhances interpretability and reasoning accuracy with less computation.

ABLE achieves state-of-the-art zero-shot reasoning performance.

Both methods effectively exploit task connections for better reasoning.

Abstract

Zero-shot event-relational reasoning is an important task in natural language processing, and existing methods jointly learn a variety of event-relational prefixes and inference-form prefixes to achieve such tasks. However, training prefixes consumes large computational resources and lacks interpretability. Additionally, learning various relational and inferential knowledge inefficiently exploits the connections between tasks. Therefore, we first propose a method for Reasoning-Oriented Locating and Editing (ROLE), which locates and edits the key modules of the language model for reasoning about event relations, enhancing interpretability and also resource-efficiently optimizing the reasoning ability. Subsequently, we propose a method for Analogy-Based Locating and Editing (ABLE), which efficiently exploits the similarities and differences between tasks to optimize the zero-shot reasoning capability. Experimental results show that ROLE improves interpretability and reasoning performance with reduced computational cost. ABLE achieves SOTA results in zero-shot reasoning.