Pronoun Resolution Neural Network Training via Iterative Feature Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current neural network models for pronoun resolution in text processing suffer from low accuracy due to direct classification methods, which fail to effectively utilize context word features and candidate substitute word features.
Innovation Solution
A data processing method and apparatus that utilize a pronoun resolution neural network to perform feature extraction, positive-example iteration processing, and negative-example iteration processing to calculate substitute probabilities, enhancing the accuracy of pronoun resolution by fusing context word and candidate substitute word features.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If direct classification method is used to resolve pronouns, then the process is simple, but the accuracy is low
Solution Approach 1:
The patent segments the pronoun resolution process into multiple stages: feature extraction from context words and candidate substitute words, positive-example iteration processing, negative-example iteration processing, and probability calculation. This segmentation allows each stage to focus on specific aspects of the problem, improving overall accuracy while maintaining manageable complexity through modular processing steps.
Solution Approach 2:
The patent transitions from direct classification to a multi-dimensional processing approach by extracting features from context words and candidate substitute words separately, then iteratively processing positive and negative examples to calculate substitute probabilities. This dimensional expansion from simple classification to multi-stage feature processing enables more accurate pronoun resolution.
2Measurement precision
If context word features and candidate substitute word features are fully utilized, then pronoun resolution accuracy improves, but data sparseness issue arises
Solution Approach 1:
The patent implements feedback mechanisms through iterative processing where positive-example iteration and negative-example iteration continuously refine the substitute probability calculations. The system uses the results from each iteration to adjust subsequent processing, effectively utilizing limited training data through feedback-driven optimization rather than relying on large quantities of data.
Solution Approach 2:
The patent changes parameters dynamically during the iterative processing stages, adjusting the weightings and thresholds for positive and negative examples based on the results from previous iterations. This parameter adaptation allows the system to maximize the utilization of available training data by optimizing parameters according to the specific characteristics of the data at hand.
Data Source
AI summary
A data processing method includes: obtaining a to-be-detected text, and determining a context word set and a candidate substitute word set corresponding to a to-be-detected word in the to-be-detected text to be inputted into a pronoun resolution neural network for feature extraction; performing positive-example iteration processing and negative-example iteration processing on the features corresponding to the context word set and the candidate substitute word set, to obtain a positive-example feature vector length and a negative-example feature vector length, and calculating a substitute probability corresponding to each candidate substitute word in the candidate substitute word set according to the positive-example feature vector length and the negative-example feature vector length; determining a target substitute word according to the substitute probability corresponding to the each candidate substitute word; and inserting the target substitute word into the to-be-detected text according to a position corresponding to the to-be-detected word, to obtain a target text.