Context-Aware Translation Quality Evaluation Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional machine learning-based translation evaluation techniques require a reference translation and often fail to account for the context of software application text, leading to inaccurate translation quality assessments.
Innovation Solution
A machine learning-based system that encodes and decodes representations of software application text and its context to generate accurate translation quality information, using a context identifier to account for the software application context, thereby reducing reliance on reference translations and improving evaluation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional machine learning-based translation evaluation techniques are used, then processing time and labor are reduced, but translation quality assessment accuracy deteriorates due to lack of context awareness
Solution Approach 1:
The patent adds a new dimension to translation evaluation by incorporating context information from software applications. Instead of evaluating translations in isolation, the system embeds translations within their original software context, allowing evaluators to assess whether translations maintain appropriate context, terminology, and usage patterns specific to the software domain.
Solution Approach 2:
The patent introduces context metadata as an intermediary element that bridges the gap between machine learning automation and accurate quality assessment. This context information acts as a mediator that provides the necessary background knowledge for evaluating translation quality without requiring manual reference translations, thereby maintaining both automation and accuracy.
2Measurement precision
If reference translations are required for evaluation, then translation quality can be assessed, but the complexity and resource requirements increase
Solution Approach 1:
The patent enables the translation evaluation system to be self-sufficient by utilizing context information from the software application itself. Instead of requiring external reference translations, the system uses the software's own context, terminology databases, and usage patterns to evaluate translation quality autonomously.
Solution Approach 2:
The patent creates a multi-functional evaluation system that can assess translation quality across different software applications and domains using a unified context-based approach. The system leverages the existing context information structure of software applications to perform evaluation without requiring application-specific customization or reference translations.
Data Source
AI summary
Systems and methods for translation evaluation are provided. One or more aspects of the systems and methods includes receiving a source text, a context identifier for the source text, and a translation text, wherein the source text comprises text from a software application and the context identifier specifies a context of the source text within the software application; generating a source text representation and a translation text representation based on the source text, the context identifier, and the translation text using an encoder of a machine learning model; and generating translation quality information based on the source text representation and the translation text representation using a decoder of the machine learning model.