Training Data Reclassification for NLU Classifier Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional techniques for generating statistical models for natural language understanding (NLU) systems do not efficiently handle outliers, leading to unreliable classification of outlier text inputs and reduced confidence scores for non-outlier classifications.
Innovation Solution
A method that processes training data by assigning sentences to fringe and core groups based on confidence scores, updating the data by associating these groups with subclasses, and building a new statistical classification model to improve classification accuracy for both outlier and non-outlier text inputs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If outliers are included within training data to adequately train the statistical model, then the model can process outlier text inputs, but the classification accuracy for non-outlier text inputs decreases and confidence scores are lower
Solution Approach 1:
The patent segments the training data into two distinct groups: outlier sentences and non-outlier sentences. This segmentation allows the statistical model to learn different patterns for each group separately, resolving the contradiction by enabling the model to handle outliers while maintaining high accuracy for non-outliers through differentiated processing
2Ease of manufacture
If conventional techniques are used to generate statistical models, then the model can be built from annotated training data, but the model cannot reliably classify outlier text inputs
Solution Approach 1:
The patent applies preliminary action by pre-processing the training data to identify and separate outlier sentences before model training. This preliminary segmentation enables the statistical model to learn from both conventional and outlier patterns effectively, improving outlier classification reliability while maintaining ease of model construction
Data Source
AI summary
A method of creating a statistical classification model for a classifier within a natural language understanding system can include processing training data using an existing statistical classification model. Sentences of the training data correctly classified into a selected class of the statistical classification model can be selected. The selected sentences of the training data can be assigned to a fringe group or a core group according to confidence score. The training data can be updated by associating the fringe group with a fringe subclass of the selected class and the core group with a core subclass of the selected class. A new statistical classification model can be built from the updated training data. The new statistical classification model can be output.