Open Input Classifier with Entailment and Pattern Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current natural language processing (NLP) systems are ineffective in narrow contexts due to the lack of vast labeled data, particularly in fields like psychological therapy, where they struggle to accurately and efficiently classify user inputs, leading to inadequate performance in providing meaningful treatment.
Innovation Solution
A computer-implemented method using a combination of entailment classifiers, pattern matching classifiers, and supervised machine learning classifiers to classify open inputs, leveraging entailment comparison data that can be updated without retraining models, facilitating accurate and efficient classification in narrow contexts such as psychological therapy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If supervised machine learning classifiers are trained on vast amounts of labeled natural language data, then classification accuracy in broad contexts is improved, but the system becomes ineffective in narrow contexts where sufficient labeled data is unavailable
Solution Approach 1:
The patent segments the classification task into multiple specialized classifiers, each trained on specific labeled data for particular contexts or domains. This allows the system to maintain high accuracy in narrow contexts by dedicating specific classifiers to those domains, rather than relying on a single general-purpose classifier that lacks sufficient specialized training data.
Solution Approach 2:
The patent creates a universal classification system that can handle both broad and narrow contexts through multiple classifiers working together. Each classifier serves specific functions for its designated context, while the ensemble provides universal coverage across different domains, making the system adaptable to both general and specialized scenarios.
2Device complexity
If a single classifier is used for classification, then the system is simpler to implement, but it cannot accurately classify inputs in narrow contexts without vast labeled data
Solution Approach 1:
The classification system is segmented into multiple specialized classifiers, each handling specific contexts or domains. This segmentation improves classification accuracy for narrow contexts by providing dedicated classifiers trained on relevant data, while maintaining manageable complexity through modular design where each classifier is relatively simple but the ensemble provides comprehensive coverage.
Data Source
AI summary
A natural language processing combination classifier is disclosed, leveraging an entailment classifier and optionally at least one of a pattern matching classifier and a trained machine learning (ML) classifier. Each of the different types of classifiers can be used to identify different categories of matches. For example, a pattern matching (e.g., regular expression) classifier may identify exact matches, an entailment classifier can obtain context-specific classifications based on likelihood of entailment to comparison data, and the SML classifier can obtain potential matches based on large-scale supervised training. The comparison data for the entailment classifier can be generated from small datasets, and can be readily updated without the need for retraining any machine learning models. Different types of classifiers can be processed using different logic to provide a user with the most appropriate response given a circumstance and given the user's open input.