CNN Relation-Entity Model for Text and Numeric Entity Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing Natural Language Processing (NLP) techniques face difficulties in accurately extracting relations between entities in text documents, particularly when numeric entities are involved, as they often fail to consider numeric inputs, leading to inaccurate results.
Innovation Solution
A method and system using a Convolutional Neural Network (CNN) model that generates entity vectors from both word and numeric embeddings, allowing for the creation of a relation-entity model to accurately determine relations between entities in text documents, including numeric inputs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If known NLP techniques use word embedding for relation extraction, then they can extract relations between text entities, but they fail to consider numeric entities leading to inaccurate results
Solution Approach 1:
The patent creates a unified embedding space that accommodates both word embeddings and numeric embeddings, allowing the same NLP model to process different entity types (text and numeric) through a single interface. This universal embedding approach enables the relation extraction system to handle diverse entity types without requiring separate processing pipelines, thereby improving adaptability while maintaining precision through consistent mathematical operations.
2Measurement precision
If supervised approaches are used with manual annotation, then relation extraction accuracy can be improved, but large volume of annotated data must be generated manually
Solution Approach 1:
The patent employs semi-supervised learning where the system automatically generates additional training examples from seed samples and unlabeled data. The model learns to identify patterns in manually annotated seed data and then applies this knowledge to automatically annotate and process large volumes of unlabeled data, significantly reducing the time and human resources required for manual annotation while maintaining high accuracy through the quality of the seed samples and the robustness of the learning algorithm.
Data Source
AI summary
A method and a system of determining a relation between two or more entities in a text document is disclosed. In an embodiment, the method may include receiving training text data annotated with two or more entities, and creating one or more n-grams based on the training text data. The method may further include generating a Convolutional Neural Network (CNN) model using the one or more n-grams, and creating an entity vector using at least one of a word embedding and a numeric embedding based on the training text data. The method may further include generating a relation-entity model using the CNN model and the entity vector.