Phrase-Level Text Entry Using Probabilistic N-gram Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current auto-complete text entry tools are limited to word-level prediction, forcing users to slow their text entry rate and interrupt their workflow to verify corrections, as they are not capable of predicting phrases, sentences, or paragraphs.
Innovation Solution
A reduced text processing module that accepts abbreviated text, parses it according to a predefined pattern, and generates probabilities for full text phrases, displaying the most likely phrases on a computing device's display component, allowing users to enter text at a higher level of abstraction such as phrases or sentences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If word-level auto-complete is used, then text entry speed is improved, but workflow interruption occurs due to verification needs
Solution Approach 1:
The patent segments text prediction from word-level to character-level, dividing the prediction task into individual character predictions that are sequentially generated. This allows the system to predict entire phrases or sentences character by character, providing more granular control and reducing verification interruptions while maintaining high speed.
Solution Approach 2:
The patent transitions from predicting complete words as discrete units to predicting individual characters as sequential elements. This dimensional change from word-level to character-level prediction enables more flexible and accurate text completion, allowing the system to adapt to various language patterns and reduce workflow interruptions.
2Measurement precision
If character-level prediction is implemented, then text entry accuracy is improved, but computational complexity increases
Solution Approach 1:
The system pre-trains character prediction models on large text corpora to learn statistical patterns and language structures. This preliminary action enables the model to make accurate character-level predictions during actual text entry with minimal computational overhead, balancing accuracy with computational efficiency.
Solution Approach 2:
The character-level prediction system automatically generates and refines predictions based on the sequence of characters already entered, without requiring manual verification or intervention. The system self-corrects and adapts to user typing patterns, maintaining high accuracy while reducing the computational burden on the user.
Data Source
AI summary
Methods, mobile electronic devices, and computer program products are provided for accepting reduced text entry of phrases, sentences or paragraphs, and probabilistically determining the most likely translation of the reduced text to a full text counterpart, and displaying same. Reduced text is accepted and parsed according to a predefined reduction pattern to produce parsed text elements. The parsed text elements are evaluated using n-gram knowledge and/or language models to identify the most likely words corresponding to the elements. The most likely corresponding words are used to evaluate the reduced text at the phrase level by evaluating the likelihood of transition from one word to the next amongst the most likely words, to compute phrase probabilities for various combinations of the most likely words. The most likely phrase(s) are output based in part on the phrase probabilities.