RecipeNet: A Hierarchical Transformer for Recipe Data

2026-08-14

AI Summary

This study addresses the limitations of existing tabular learning methods in capturing hierarchical field interactions and process dependencies within recipe data. To overcome these challenges, this work proposes RecipeNet, which employs a hierarchical Transformer architecture. By stacking encoders to jointly model intra-step field interactions and inter-step sequential dependencies, the method effectively integrates both structured and sequential characteristics of recipes. Experimental results demonstrate that RecipeNet consistently outperforms state-of-the-art tabular models across multiple benchmark datasets. These findings validate the efficacy of hierarchical sequence modeling for complex recipe representation learning and establish a novel paradigm for industrial recipe data mining.

Abstract

Recipe data arises in domains such as materials synthesis, pharmaceutical formulation, and industrial manufacturing, where procedures are represented as ordered sequences of steps containing heterogeneous structured fields. Existing tabular learning methods typically flatten this structure into fixed-schema representations, limiting their ability to capture hierarchical field interactions and procedural dependencies. We propose RecipeNet, a hierarchical Transformer architecture that encodes field-level interactions within each step and sequential dependencies across steps through stacked Transformer encoders. Experiments on multiple recipe datasets and tasks demonstrate that RecipeNet consistently outperforms existing tabular models, highlighting the value of hierarchical and sequential modeling for recipe representation learning.

Problem

• Recipe data
• Tabular learning
• Hierarchical field interactions
• Procedural dependencies
• Representation learning

Innovation

• Hierarchical Transformer
• Recipe Representation Learning
• Sequential Dependencies
• Field-level Interactions
• Structured Data