Phonemic Text Correction via TTS-STS Conversion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Standard English language processing services face difficulties in handling phonemic spellings, which are considered erroneous and do not map to standard English words, leading to inaccuracies in processing messages, especially from non-native speakers and social media texts.
Innovation Solution
A computer-implemented method and system that uses sequential text-to-speech and speech-to-text recognition algorithms to convert phonemic spellings into correct graphemic representations by identifying the primary language and accent of the writer, utilizing APIs like Watson from IBM and Amazon Polly, to generate an audio file and subsequently correct the spelling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If standard English language processing services are used to process phonemic spellings, then processing speed is maintained, but processing accuracy deteriorates because phonemic words are considered erroneous and do not map to standard English words
Solution Approach 1:
The patent introduces an intermediary processing system that bridges phonemic spellings and standard English processing. The system includes: (1) a phoneme-to-grapheme conversion module that converts phonemic spellings to standard spelling, (2) a language model that predicts intended words from phonemic sequences, and (3) an integration layer that connects to standard NLP services. This intermediary approach maintains compatibility with existing services while accurately handling phonemic variations, thereby improving processing accuracy without requiring complete system redesign.
Solution Approach 2:
The patent changes the parameter representation of text from strict grapheme-based encoding to a hybrid phoneme-grapheme representation. Specifically, it implements: (1) phoneme segmentation that identifies phonemic spellings, (2) dynamic parameter adjustment that switches between phonemic and standard processing modes, and (3) adaptive mapping that transforms phonemic parameters to standard English parameters. This parameter transformation enables accurate processing of phonemic text while maintaining compatibility with standard processing pipelines.
2Productivity
If phonemic spellings are used in text messages, then typing speed increases, but language processing accuracy deteriorates because standard services cannot recognize these spellings
Solution Approach 1:
The patent applies preliminary action by performing phoneme-to-grapheme conversion and spelling correction before the text enters standard language processing services. The system: (1) detects phonemic spellings in advance, (2) converts them to standard spelling patterns, and (3) validates the converted text against a language model. This preliminary processing ensures that by the time text reaches standard NLP services, it is in the correct format, thereby maintaining both typing speed benefits and processing accuracy.
Solution Approach 2:
The patent implements feedback mechanisms that monitor the effectiveness of phonemic spelling conversion. The system includes: (1) accuracy validation that checks converted text against language models, (2) error detection that identifies failed conversions, and (3) adaptive learning that adjusts conversion parameters based on feedback. This feedback loop ensures high accuracy in converting phonemic spellings while preserving the productivity benefits of rapid typing.
3Loss of information
If phonemic orthography is used to write English, then representation of spoken sounds improves, but compatibility with standard English processing services deteriorates
Solution Approach 1:
The patent adds a dimensional transformation layer that converts phonemic text into standard English representation. The system operates in multiple dimensions: (1) phonemic dimension that preserves spoken sound information, (2) graphemic dimension that provides standard spelling, and (3) semantic dimension that ensures meaningful interpretation. By transforming text across these dimensions, the system maintains phonetic accuracy while achieving compatibility with standard processing services that operate in the graphemic dimension.
Solution Approach 2:
The patent creates a universal processing framework that handles both phonemic and standard English text through a single system. The framework includes: (1) a unified conversion engine that processes both phonemic and standard input, (2) adaptive service interfaces that work with multiple processing services, and (3) cross-format compatibility layers. This universal approach allows the system to maintain phonetic accuracy for phonemic input while achieving broad compatibility with standard English processing services.
Data Source
AI summary
Correcting typographical errors in electronic text may include converting a text message containing at least one phonemic spelling of a word into speech by running a text-to-speech application programming interface (API) with the text message as input. The converted speech may be input to a speech-to-text API and the speech-to-text API executed to convert the speech to text. A text file comprising the text may be generated and/or output. The text file automatically contains a corrected version of the phonemic spelling of the word in text message.