Recipe Ingredient Parsing for Food Item Action Pair Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In the cooking domain, creating a model for action/object compatibility is challenging due to syntactical inconsistencies in recipe text, making it difficult to automate the process of learning verb/noun collocations and generating rules for recipe ingredients.
Innovation Solution
A method is provided to extract ingredient lists from recipes, identify actions associated with each ingredient by parsing multi-word entries, and create a lexicon that associates food items with actions, using a verb lexicon and clause structure analyzer to generate food item/action pairs, which can be used by recipe generators like Chef Watson.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual model creation is used for action/object compatibility, then model accuracy can be ensured, but the process becomes very time consuming
Solution Approach 1:
The system performs self-learning by automatically parsing recipe texts and extracting verb/noun collocations without requiring manual annotation of each recipe. The learning engine autonomously processes the corpus of recipes to build the action/object compatibility model, eliminating the need for manual model creation while maintaining reliability through systematic analysis of cooking domain terminology and syntactic patterns.
2Productivity
If learning from data corpus is used, then the process is faster and more reliable, but syntactical inconsistencies make it difficult to figure out dividing lines for generating rules
Solution Approach 1:
The system changes the parameter of text representation by converting unstructured recipe text into structured parse trees using grammatical analysis. This transformation reorganizes the syntactical data into a standardized format where verb/noun relationships are clearly defined by grammatical roles, making it easier to extract collocations despite the original syntactical inconsistencies in the recipe corpus.
Solution Approach 2:
The patent introduces an intermediary learning engine that acts as a mediator between the raw recipe corpus and the final action/object model. This intermediary layer performs grammatical parsing and applies cooking domain-specific rules to bridge the gap between inconsistent syntactical forms and consistent semantic relationships, systematically handling the complexity of extracting meaningful collocations from varied recipe expressions.
3Ease of operation
If standard parsing is used on recipe text, then general text processing is simplified, but parsing errors increase due to non-standard recipe language
Solution Approach 1:
The system applies local quality by using domain-specific grammatical rules tailored to recipe text rather than general-purpose parsing. The learning engine incorporates cooking domain knowledge to handle specific syntactical patterns common in recipes, such as ingredient lists with implicit verbs and instructional steps with imperative mood, thereby improving parsing accuracy for this specific text type while maintaining the benefits of automated processing.
Data Source
AI summary
Aspects include learning object/action pairs for recipe ingredients. A method includes extracting an ingredient list that includes a plurality of entries from a recipe stored in an electronic format. Each of the plurality of entries includes one or more words. For each multi-word entry in the ingredient list, an ingredient associated with the entry is identified by comparing words contained in the entry to terms in a lexicon of cooking domain terms. It is determined whether an action corresponds to the identified ingredient. The determining includes parsing the entry to locate a verb that is associated with the identified ingredient. Based on determining that an action corresponds to the identified ingredient, a food item action pair that includes the identified ingredient and the located verb is output.