Translation Memory Sub-segment Extraction for Reuse

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current natural language translation methods face limitations in reusing high-quality previously translated text due to their high granularity, requiring significant human effort for term base development and maintenance, and lacking efficient integration of previously translated text in machine-assisted translation projects.

Innovation Solution

A computer-implemented method that extracts and suggests target text sub-segments from a corpus of previously translated text pairs, allowing for the reuse of high-quality translations at a lower level of granularity, such as words or phrases, thereby reducing human input and increasing translation efficiency by providing relevant suggestions to the translator.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If translation memories leverage existing translations on the sentence or paragraph level, then the translation quality is maintained, but the amount of re-use possible is limited due to the relatively low chance of a whole sentence or paragraph matching the source text

Engineering Contradiction:

Improveamount of re-use of previously translated textVSAvoidgranularity of text segments

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent divides previously translated text segments into smaller sub-segments (words or phrases) rather than using whole sentences or paragraphs. This segmentation increases the granularity control, allowing the system to match and reuse appropriate sized portions of translated text that better fit the source text being translated, thereby increasing the amount of reusable content.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system extracts relevant sub-segments from previously translated text segments based on the current translation context. By extracting only the necessary portions (words or phrases) that match the source text, the system enables more flexible and frequent reuse of translated content without requiring complete sentence or paragraph matches.

Inventive Principle:

Principle #2Taking out (Extraction)

2Productivity

If a term base or multilingual dictionary is built up from previous translations, then the leverage of previous translations is improved, but the human effort required in creating and maintaining such term bases is considerable

Engineering Contradiction:

Improveleverage of previous translationsVSAvoidhuman effort for term base development and maintenance

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The system automatically extracts term candidates from existing monolingual or bilingual resources without requiring manual intervention. The extraction process is performed autonomously by the computer system, identifying and organizing relevant terms and translations from previously translated text, thereby eliminating the need for skilled terminologists to manually create and maintain term bases.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces the manual mechanical process of term base creation and maintenance with an automated computational system. The system uses algorithms to automatically extract, organize, and manage translation terms and sub-segments from existing translation memories, substituting human effort with automated information processing.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If word processors use internal heuristics to suggest potential completions of a typed-in prefix, then the typing speed is improved, but no previously translated text is re-used in machine translation capabilities

Engineering Contradiction:

Improvetyping speed and auto-completionVSAvoidre-use of previously translated text

Core Design Contradiction:

ProductivityVSLoss of information

Solution Approach 1:

The system acts as an intermediary between the translation memory database and the translator's interface. It retrieves previously translated sub-segments from the translation memory based on the current translation context and presents them as suggestions to the translator, combining the benefits of auto-completion with the reuse of professionally translated content.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system provides feedback to the translator by suggesting relevant previously translated sub-segments during the translation process. This feedback mechanism allows translators to see and select from proven translations, improving both speed and quality by leveraging existing translation knowledge rather than relying solely on generic auto-completion algorithms.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS8935148B2Computer-assisted natural language translation

Publication Date: 2015.01.13 SDL LTD

AI summary

A computer implemented method of translating source material in a source natural language into a target natural language includes receiving a first data input which is a first part of a sub-segment of a translation of the source material from the source natural language into the target natural language, identifying a selectable target text sub-segment in the target natural language associated with the received first data input, and outputting the selectable target text sub-segment. The selectable target text sub-segment is extracted from a corpus of previously translated text segment pairs, each text segment pair having a source text segment in the source natural language and a corresponding translated text segment in the target natural language.