Chapter-Level Text Translation Model for Contextual Consistency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional text translation methods struggle with contextual consistency and pronoun supplementation across sentences, leading to inaccuracies in document, novel, and video translation.
Innovation Solution
A method and apparatus for text translation that determine a keyword set including entity words and pronouns from a chapter-level monolingual corpus, mask the corpus based on this set, and generate a chapter-level text translation model to improve translation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional sentence-by-sentence translation is used, then translation speed is improved, but translation accuracy and contextual consistency deteriorate
Solution Approach 1:
The patent segments the translation task into two distinct stages: a sentence-level translation stage for maintaining translation speed, and a chapter-level refinement stage for improving accuracy and contextual consistency. This segmentation allows the system to benefit from both fast sentence-level processing and accurate context-aware refinement.
Solution Approach 2:
The patent performs preliminary sentence-level translation to establish a baseline translation quickly, then applies subsequent chapter-level refinement using monolingual corpus masking and entity-pronoun relationship learning. This preliminary action ensures that the translation process starts with speed and then optimizes for accuracy.
2Device complexity
If sentence-by-sentence translation is used, then processing simplicity is improved, but contextual semantic relationship preservation deteriorates
Solution Approach 1:
The patent divides the translation process into simple sentence-level processing for initial translation and a more complex chapter-level refinement stage that handles contextual semantic relationships. This segmentation maintains overall processing simplicity while addressing contextual consistency through targeted refinement.
Solution Approach 2:
The patent introduces a chapter-level monolingual corpus as an intermediary resource that contains masked sentences with entity-pronoun relationship information. This intermediary enables the system to learn and preserve contextual semantic relationships without complicating the basic sentence-level translation process.
3Measurement precision
If monolingual corpus masking with keyword sets is applied, then translation model accuracy is improved, but training complexity increases
Solution Approach 1:
The patent performs preliminary processing of the monolingual corpus to extract keyword sets including entity words and pronouns before applying the masking technique. This preliminary action organizes the training data structure, making the subsequent masking and model training process more manageable and less complex.
Solution Approach 2:
The patent applies masking selectively to specific keyword sets (entity words and pronouns) within the monolingual corpus rather than uniformly masking all words. This local quality approach focuses the training complexity on the most important contextual elements while keeping the overall training process efficient.
Data Source
Figure 1
Figure 2A~2B
Figure 2C~2D
AI summary
Embodiments of the present disclosure relate to a method, an apparatus, an electronic device, and a medium for text translation. The method includes determining a keyword set associated with a chapter-level monolingual corpus in a target language, where the keyword set includes a plurality of entity words and a plurality of pronouns, and masking the chapter-level monolingual corpus based on the keyword set. The method further includes generating a chapter-level text translation model based on the masked chapter-level monolingual corpus. According to the embodiments of the present disclosure, it is possible to enable translations of the same or associated words to have contextual consistency throughout a text, and to explicit a noun indicated by a pronoun, and further to supplement a missing pronoun, thereby improving accuracy of the text translation model.