Word Relationship Discrimination Accuracy via Prediction Model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to enhance the discrimination accuracy of models that extract relationships between words from sentences, particularly in the chemistry field, due to the difficulty in preparing large amounts of appropriate learning data.
Innovation Solution
An information processing method that uses a prediction model to specify words appearing before or after a target word in a sentence, and then determines the appropriateness of the estimated relationships based on stored rules regarding word units, to generate high-quality learning data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a prediction model is used to specify words before or after a target word, then the discrimination accuracy of word relationship models is improved, but the complexity of the processing system increases
Solution Approach 1:
The prediction model performs preliminary action by predicting words that appear before or after a target word in advance. This preliminary specification of context words enables the subsequent relationship estimation process to operate with pre-identified candidates, improving discrimination accuracy while managing system complexity through proactive data preparation.
Solution Approach 2:
The prediction model acts as an intermediary component between the input sentence and the relationship estimation process. It mediates by generating predicted context words that bridge the gap between raw text and relationship analysis, thereby improving the accuracy of word relationship discrimination without requiring direct complex interactions between all sentence components.
2Reliability
If rules regarding word units are used to determine appropriateness of relationships, then the quality of learning data is improved, but the time required for relationship verification increases
Solution Approach 1:
The system implements feedback by using stored rules regarding word units to verify and validate the estimated relationships between words. This feedback mechanism ensures that only relationships conforming to established linguistic patterns are included in the learning data, improving data quality while the automated nature of the verification minimizes time loss through efficient rule-based checking.
Data Source
AI summary
An information processing method in which a computer executes processing includes: acquiring a sentence; specifying a word that appears immediately before or immediately after a first word in the acquired sentence by using a prediction model that predicts a word that appears immediately before or immediately after an input word; determining whether or not an estimated relationship between the first word and the second word in the sentence is appropriate on the basis of the specified word and a rule regarding a unit that corresponds to a relationship between words stored in a storage; and outputting information regarding the estimated relationship between the first word and the second word in a case where it is determined that the relationship is appropriate.