Translated Clause Generation via Conjunction-Based Corpus Construction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current machine translation systems face difficulties in translating long clauses or complex sentences effectively, particularly when multiple clauses are joined together with conjunctions, as they lack adequate bilingual corpora containing such structures.
Innovation Solution
A method and apparatus for generating translated clauses by acquiring and translating individual clauses, then constructing them into a single sentence using appropriate conjunctions in both the source and target languages, thereby improving the translation of long clauses by creating a bilingual corpus that includes joined clauses with conjunctions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If machine translation systems translate long clauses or complex sentences directly, then translation coverage is improved, but translation accuracy deteriorates
Solution Approach 1:
The patent segments long clauses or complex sentences into multiple shorter clauses before translation. The translation processing unit divides the input long clause into several short clauses, translates each separately using the translation table, and then reconstructs them into the final translated sentence. This segmentation approach enables the system to handle diverse translation scenarios (improving adaptability) while maintaining high accuracy for each translated segment through proven translation patterns.
2Ease of manufacture
If bilingual corpus contains only simple clauses, then corpus creation is easier, but translation performance of complex sentences deteriorates
Solution Approach 1:
The patent performs preliminary action by pre-processing long clauses into segmented short clauses before translation. The translation table stores translation patterns for these segmented clauses, and the processing unit applies these pre-prepared patterns to reconstruct accurate translations of complex sentences. This preliminary segmentation and pattern storage enables the corpus to handle complex sentences effectively without requiring exhaustive manual creation of all possible complex sentence pairs.
3Measurement precision
If translation system uses segmented clause approach, then translation accuracy of long clauses is improved, but processing complexity increases
Solution Approach 1:
The patent uses copying by storing translation patterns of segmented clauses in a translation table during the corpus creation phase. During actual translation processing, the system copies these pre-stored patterns from the translation table to generate translations, rather than performing complex analysis each time. This copying mechanism maintains high translation accuracy for long clauses while significantly reducing the processing complexity during runtime operations.
Data Source
AI summary
A bilingual-corpus generating method includes: acquiring a first clause in the first language, a first translated clause obtained by translating the first clause into a second language, a second clause in the first language, a second translated clause obtained by translating the second clause into the second language; and generating, as a translated clause of a third clause obtained by constructing the first clause and the second clause as one sentence by using a first conjunction in the first language, a third translated clause obtained by constructing the first translated clause and the second translated clause as one sentence by using a second conjunction in the second language.