Chinese Speech Punctuation Accuracy via Semantic Feature Integration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for adding punctuation to speech documents, such as relying on mute interval thresholds or hyphenation processing, result in low accuracy due to limited information usage and poor association between character positions and punctuation status.
Innovation Solution
A computer-implemented method that processes information content by identifying semantic features and characteristic units, extracting probabilities for punctuation marks, and establishing a language model based on frequency occurrences to accurately determine punctuation marks for unsegmented content.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If mute interval threshold is used to add punctuation, then the process is simple, but the accuracy of punctuation addition is low
Solution Approach 1:
The patent combines multiple features (character position information, semantic features, punctuation type, and context) into a unified language model. This merging of multiple information sources resolves the contradiction by maintaining process simplicity while significantly improving punctuation accuracy through comprehensive feature integration.
Solution Approach 2:
The patent transforms the single parameter approach (mute interval threshold) into a multi-parameter language model that considers character position, semantic features, punctuation types, and contextual information. This parameter transformation resolves the contradiction by enabling accurate punctuation addition without increasing operational complexity.
2Loss of information
If hyphenation processing and character position are used to establish language model, then more information is utilized, but the association between character position and punctuation status remains poor
Solution Approach 1:
The patent introduces semantic features as an additional dimension to the language model, transforming the two-dimensional character position-punctuation association into a multi-dimensional space that includes semantic context. This dimensional expansion resolves the contradiction by enabling stronger associations between character positions and punctuation status through enriched feature representation.
3Adaptability or versatility
If speaking speed varies, then natural speech is captured, but punctuation addition accuracy deteriorates due to threshold mismatch
Solution Approach 1:
The patent implements a dynamic language model that adapts to varying speaking speeds by incorporating contextual information and semantic features. This dynamic approach resolves the contradiction by enabling the system to maintain high punctuation accuracy across different speaking speeds through flexible feature weighting and context-aware decision making.
Data Source
AI summary
A method of processing information content based on a Chinese language model is performed at a computer, the method including: identifying a plurality of expressions in the information content extracted from a speech input through speech recognition that is queued to be processed; dividing the expressions into a plurality of characteristic units according to semantic features and predetermined characteristics associated with each characteristic unit, each including a subset of the expressions and the predetermined characteristics at least including a respective integer number of expressions that are included in the characteristic unit; extracting, from the Chinese language model, a plurality of probabilities for punctuation marks associated with each characteristic unit; and in accordance with the probabilities, associating a respective punctuation mark with each characteristic unit included in the information content. The method further comprises adding punctuation marks based on a weight determined for each punctuation mark.