Morpheme-Level Predictive Keyboard Input Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing computing devices with word-level prediction techniques require a large number of user inputs to enter long words, as they predict entire words rather than their constituent morphemes, leading to inefficiencies in text input processes.
Innovation Solution
Implementing morpheme-level word prediction techniques that break down words into their constituent morphemes, allowing for more fine-grained predictions and optimizing candidate word selection using algorithms that reduce the number of required user inputs by presenting candidate words with morpheme-level divisions and optimizing at the morpheme level.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If word-level prediction techniques are used, then the system can predict complete words, but the number of user inputs required increases for long words
Solution Approach 1:
The patent segments words into morphemes (smallest meaningful units) and performs prediction at the morpheme level rather than complete word level. This allows the system to predict words like 'unbelievable' by predicting morphemes 'un-', 'believ-', '-able' separately, reducing the number of user inputs from 11 characters to approximately 3-4 morpheme selections while maintaining high prediction accuracy
Solution Approach 2:
The patent introduces a new dimension of prediction granularity by operating at the morpheme level between character-level and word-level prediction. This intermediate dimension allows for more efficient prediction of long words while maintaining the benefits of word-level completion, effectively adding a temporal efficiency dimension to the prediction process
2Adaptability or versatility
If word-level prediction is used, then candidate words can be generated, but the number of candidate words remains large for long words requiring many inputs
Solution Approach 1:
By segmenting the prediction task into morpheme-level units, the system reduces the search space from all possible long words to morphemes that can be combined. This segmentation allows the prediction algorithm to work with smaller, more manageable units while still generating diverse and accurate candidate words through morpheme combination
Solution Approach 2:
The system performs partial prediction by completing only the necessary morphemes rather than requiring full word input. This partial action approach predicts words with fewer inputs by identifying and predicting only the morphemic structure needed to form valid words, reducing system complexity while maintaining versatility
Data Source
AI summary
In one example, a method includes determining, by a computing device and based at least in part on an initial character string, one or more candidate morpheme sequences, wherein each of the candidate morpheme sequences includes the initial character string and one or more candidate morphemes. The method further includes outputting, for display, the one or more candidate morpheme sequences. The method further includes receiving an indication of a user input detected at a presence-sensitive input device. The method further includes selecting, based on the indication of the user input, at least one of the candidate morphemes from one of the candidate morpheme sequences to define a selected morpheme sequence that includes the initial character string and the selected candidate morpheme from the one of the candidate morpheme sequences. The method further includes outputting, for display, the selected morpheme sequence.