Automatic Extraction of Relationships among Motivations, Emotions and Actions from Natural Language Texts

TL;DR

This paper introduces a graph-based framework leveraging large language models to extract and analyze relationships among motivations, emotions, and actions from natural language texts, specifically focusing on food reviews.

Contribution

It presents a novel, annotation-free method using large language models to generate and analyze relationship graphs in natural language texts.

Findings

Generated 92,990 relationship graphs from food reviews.

63% of the graphs are logically consistent.

Provides insights into error types for future improvements.

Abstract

We propose a new graph-based framework to reveal relationships among motivations, emotions and actions explicitly given natural language texts. A directed acyclic graph is designed to describe human's nature. Nurture beliefs are incorporated to connect outside events and the human's nature graph. No annotation resources are required due to the power of large language models. Amazon Fine Foods Reviews dataset is used as corpus and food-related motivations are focused. Totally 92,990 relationship graphs are generated, of which 63% make logical sense. We make further analysis to investigate error types for optimization direction in future research.