Electronic Device User Dictionary Semantic Autocomplete
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing word prediction systems in smart devices, such as those using neural network-based language models, fail to provide user-specific words as automatic completion recommendations, even if they have similar meanings to learned words, due to their limited vocabulary and lack of semantic analysis for user-specific terms.
Innovation Solution
An electronic device with a user-based dictionary that receives input sentences containing user-specific words and learned words, determines the concept category of the user-specific word based on semantic information, adds it to the dictionary, and provides it as an automatic completion recommendation when relevant, thereby enhancing input functionality and personalization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If semantic word embedding is used in neural network language model, then semantic consistency of word recommendations is improved, but user-specific words cannot be recognized and recommended
Solution Approach 1:
The patent segments the language model into two parts: a neural network language model for semantic word embedding and a user-based dictionary for user-specific words. This segmentation allows each component to specialize - the neural network handles semantic consistency while the user-based dictionary handles user-specific word recognition, resolving the contradiction between these two requirements.
Solution Approach 2:
The user-based dictionary acts as an intermediary between the neural network language model and user-specific words. It stores user-specific words and their semantic information, enabling the neural network to recognize and recommend user-specific words while maintaining semantic consistency through the mediation of the dictionary.
2Ease of manufacture
If N-gram model is used for word prediction, then ease of adding words is improved, but semantic consistency of recommendations deteriorates
Solution Approach 1:
The patent merges the N-gram model's ease of word addition capability with the neural network's semantic word embedding capability. The system combines both approaches, allowing words to be easily added to the user-based dictionary while maintaining semantic consistency through the neural network's word embedding and semantic analysis.
3Measurement precision
If neural network language model is used, then semantic analysis capability is improved, but vocabulary coverage is limited
Solution Approach 1:
The user-based dictionary performs preliminary action by storing user-specific words and their semantic information before they are needed for prediction. This preliminary preparation allows the neural network to recognize and analyze user-specific words without limiting vocabulary coverage, as the dictionary pre-stores the expanded vocabulary with semantic annotations.
Data Source
AI summary
Provided are an electronic device and a control method. The electronic device comprises: a storage unit for storing a user-based dictionary; an input unit for receiving an input sentence including a user-specific word and at least one word learned by a neural network-based language model; and a processor for determining a concept category of the user-specific word on the basis of semantic information of the input sentence, adding the user-specific word to the user-based dictionary to perform update, and when text corresponding to semantic information of the at least one learned word is input, providing the user-specific word as an autocomplete recommendation word which can be input subsequent to the text.