Unsupervised Learning of Graph from Recipes

TL;DR

This paper introduces an unsupervised method to extract and represent the sequence of actions in cooking recipes as graphs using a graph neural network, without relying on labeled data.

Contribution

It presents a novel unsupervised approach that jointly learns recipe graphs and text encoding, outperforming existing methods in entity identification and graph accuracy.

Findings

Effective entity extraction from recipes

Accurate graph generation compared to state-of-the-art methods

Improved text reconstruction quality

Abstract

Cooking recipes are one of the most readily available kinds of procedural text. They consist of natural language instructions that can be challenging to interpret. In this paper, we propose a model to identify relevant information from recipes and generate a graph to represent the sequence of actions in the recipe. In contrast with other approaches, we use an unsupervised approach. We iteratively learn the graph structure and the parameters of a $\mathsf{GNN}$ encoding the texts (text-to-graph) one sequence at a time while providing the supervision by decoding the graph into text (graph-to-text) and comparing the generated text to the input. We evaluate the approach by comparing the identified entities with annotated datasets, comparing the difference between the input and output texts, and comparing our generated graphs with those generated by state of the art methods.