Retranslating Queries via Semantic Matching for Translation Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing conversation systems fail to accurately translate query texts, often resulting in responses that do not match the intended query intention, leading to miscommunication between users of different languages.
Innovation Solution
An electronic apparatus and method that performs retranslation of a query text based on a response text by using neural network models to identify semantic matches and adjust the query accordingly, including sentence division and paraphrasing models to refine the translation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a translation model translates a first language into a second language, then language communication between users is enabled, but translation accuracy deteriorates causing responses to not match query intention
Solution Approach 1:
The system uses a semantic matching model to evaluate whether the response text semantically matches the query text, and feeds this evaluation back to determine if retranslation is needed. This feedback loop enables the system to identify and correct translation inaccuracies while maintaining multilingual communication capability.
Solution Approach 2:
The system performs preliminary semantic matching between the query text and response text before finalizing the translation output. By checking semantic consistency in advance, the system can identify translation errors early and trigger retranslation to improve accuracy while preserving language communication functionality.
2Measurement precision
If retranslation is performed to improve translation accuracy, then semantic matching between query and response improves, but system complexity increases due to additional neural network models
Solution Approach 1:
The semantic matching model serves multiple functions: it evaluates semantic consistency between query and response, determines whether retranslation is needed, and can guide the retranslation process. This multi-functionality reduces the need for separate dedicated components for each task, thereby managing system complexity while improving translation accuracy.
Solution Approach 2:
The system uses its own semantic matching capability to automatically detect translation errors and trigger retranslation without external intervention. The translation system self-corrects by identifying mismatches and performing retranslation when needed, reducing the need for complex external validation systems.
3Measurement precision
If semantic matching is performed using neural network models, then translation quality improves, but processing time increases
Solution Approach 1:
The system performs semantic matching selectively rather than for every translation. It uses the semantic matching model to evaluate translations only when needed to improve quality, avoiding the time cost of universal semantic analysis while maintaining high translation quality for critical communications.
Data Source
AI summary
A method for controlling an electronic apparatus includes: translating a first query text of a first language to acquire a second query text of a second language; transmitting the second query text to an external device; acquiring, from the external device, a first response text of the second language in response to the second query text; acquiring a second response text acquired by translating the first response text into the first language, and identifying whether the second response text semantically matches to the first query text by inputting the second response text and the first query text into a first neural network model configured to identify whether a query and a response semantically match; and acquiring a third query text of the second language by retranslating the first query text based on a result of identifying that the first query text and the second response text do not semantically match.