Food Recipe Alternation and Generation with Natural Language Processing Techniques

Pan, Yuran; Xu, Qiangwen; Li, Yanjun

doi:10.1109/icdew49219.2020.000-1

Cited by 14 publications

(7 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Akkoyunlu et al (2017) similarly used the context of foods consumed together and applied a penalty to food items frequently consumed together. The use of NLP techniques and embedding similarity to search for substitute ingredients was explored by Pan et al (2020) , but this work lacked a formal evaluation of the generated substitutions. Our work differs from such previous works in that we leverage a greater degree of explicit semantic information about foods.…”

Section: Discussionmentioning

confidence: 99%

“…To compare the performance of our method against other recent works in the domain of food, we additionally develop a Poincaré embedding model (Nickel and Kiela, 2017) using FoodOn's class hierarchy. FoodEx2Vec (Eftimov et al, 2020) demonstrated the utility of Poincaré graph embeddings, which can capture hierarchical relations between terms, to develop embedding for a food classification system.…”

Section: Poincaré Embeddingsmentioning

confidence: 99%

See 1 more Smart Citation

Identifying Ingredient Substitutions Using a Knowledge Graph of Food

Shirai

Seneviratne

Gordon

et al. 2021

Front. Artif. Intell.

View full text Add to dashboard Cite

People can affect change in their eating patterns by substituting ingredients in recipes. Such substitutions may be motivated by specific goals, like modifying the intake of a specific nutrient or avoiding a particular category of ingredients. Determining how to modify a recipe can be difficult because people need to 1) identify which ingredients can act as valid replacements for the original and 2) figure out whether the substitution is “good” for their particular context, which may consider factors such as allergies, nutritional contents of individual ingredients, and other dietary restrictions. We propose an approach to leverage both explicit semantic information about ingredients, encapsulated in a knowledge graph of food, and implicit semantics, captured through word embeddings, to develop a substitutability heuristic to rank plausible substitute options automatically. Our proposed system also helps determine which ingredient substitution options are “healthy” using nutritional information and food classification constraints. We evaluate our substitutability heuristic, diet-improvement ingredient substitutability heuristic (DIISH), using a dataset of ground-truth substitutions scraped from ingredient substitution guides and user reviews of recipes, demonstrating that our approach can help reduce the human effort required to make recipes more suitable for specific dietary needs.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Poincaré Embeddingsmentioning

confidence: 99%

Identifying Ingredient Substitutions Using a Knowledge Graph of Food

Shirai

Seneviratne

Gordon

et al. 2021

Front. Artif. Intell.

View full text Add to dashboard Cite

show abstract

“…Previous methods to edit recipes focused on broad classes like dietary categories [12] and cuisines [16] and require paired corpora (which do not exist for fine-grained edits). We propose a method that does not require paired corpora to train and accommodates positive and negative user feedback on an ingredient-level.…”

Section: Recipecrit: a Hierarchical Denoising Recipe Auto-encodermentioning

confidence: 99%

“…Recipe editing can be seen as a combination of recipe generation and controllable natural language generation [22]. It has recently been explored for creating recipes that satisfy a specific dietary constraint [12] or follow a specified cuisine [16]. On one hand, pretrained language models have been used to create recipe directions given a known title and set of ingredients [4,9,10], but generated recipes suffer from inconsistency [12].…”

Section: Introductionmentioning

confidence: 99%

Assistive Recipe Editing through Critiquing

Antognini¹,

Li²,

Faltings³

et al. 2022

Preprint

View full text Add to dashboard Cite

There has recently been growing interest in the automatic generation of cooking recipes that satisfy some form of dietary restrictions, thanks in part to the availability of online recipe data. Prior studies have used pre-trained language models, or relied on small paired recipe data (e.g., a recipe paired with a similar one that satisfies a dietary constraint). However, pre-trained language models generate inconsistent or incoherent recipes, and paired datasets are not available at scale. We address these deficiencies with RecipeCrit, a hierarchical denoising auto-encoder that edits recipes given ingredientlevel critiques. The model is trained for recipe completion to learn semantic relationships within recipes. Our work's main innovation is our unsupervised critiquing module that allows users to edit recipes by interacting with the predicted ingredients; the system iteratively rewrites recipes to satisfy users' feedback. Experiments on the Recipe1M recipe dataset show that our model can more effectively edit recipes compared to strong language-modeling baselines, creating recipes that satisfy user constraints and are more correct, serendipitous, coherent, and relevant as measured by human judges.

show abstract

“…However, the current works on using AI for food recipe ingredient substitution are, so far, very scarce. Pan et al [6] examined how to use natural language processing techniques, such as word embeddings, to find alternative components in a data-driven, similarity-based manner. Several systems, on the other hand, incorporated explicit semantic information about ingredients and explicit rules.…”

Section: Introductionmentioning

confidence: 99%