Enregistré dans:
Détails bibliographiques
Auteurs principaux: Radzikowski, Jakub, Chen, Josef
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:https://arxiv.org/abs/2605.22391
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866911704622825472
author Radzikowski, Jakub
Chen, Josef
author_facet Radzikowski, Jakub
Chen, Josef
contents We present Epicure, a family of three sibling skip-gram ingredient embeddings retrained from scratch on a multilingual recipe corpus. We aggregate 4.14M recipes from 11 sources spanning seven languages, English, Chinese, Russian, Vietnamese, Spanish, Turkish, Indonesian, German, and Indian-English, and normalise the raw ingredient strings to 1,790 canonical entries via an LLM-augmented pipeline. A 203,508-edge ingredient-ingredient NPMI graph and an 80,019-edge typed FlavorDB ingredient-compound graph, 2,247 typed compound nodes across 15 categories, seed three Metapath2Vec variants that share architecture and hyperparameters and differ only in the random-walk schema: Cooc walks the co-occurrence graph only, Chem walks the typed compound metapaths only, and Core blends both via injected ingredient-ingredient walks at controlled mixing, placing each model at a distinct point on the chemistry-vs-recipe-context spectrum.
format Preprint
id arxiv_https___arxiv_org_abs_2605_22391
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Epicure: Navigating the Emergent Geometry of Food Ingredient Embeddings
Radzikowski, Jakub
Chen, Josef
Artificial Intelligence
Computation and Language
Computers and Society
We present Epicure, a family of three sibling skip-gram ingredient embeddings retrained from scratch on a multilingual recipe corpus. We aggregate 4.14M recipes from 11 sources spanning seven languages, English, Chinese, Russian, Vietnamese, Spanish, Turkish, Indonesian, German, and Indian-English, and normalise the raw ingredient strings to 1,790 canonical entries via an LLM-augmented pipeline. A 203,508-edge ingredient-ingredient NPMI graph and an 80,019-edge typed FlavorDB ingredient-compound graph, 2,247 typed compound nodes across 15 categories, seed three Metapath2Vec variants that share architecture and hyperparameters and differ only in the random-walk schema: Cooc walks the co-occurrence graph only, Chem walks the typed compound metapaths only, and Core blends both via injected ingredient-ingredient walks at controlled mixing, placing each model at a distinct point on the chemistry-vs-recipe-context spectrum.
title Epicure: Navigating the Emergent Geometry of Food Ingredient Embeddings
topic Artificial Intelligence
Computation and Language
Computers and Society
url https://arxiv.org/abs/2605.22391