Context-Based Phrase Generation for Translation Accuracy

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional machine translation methods fail to accurately convey the meaning and intent of words or phrases, especially idiomatic expressions, when translating text from one language to another, leading to reader confusion due to word-by-word translation that loses contextual information.

Innovation Solution

A computer system monitors user interaction to detect difficulty with phrases and generates semantically similar alternative phrases by analyzing context, tone, and using a monolingual corpus to provide alternative translations that capture the intended meaning, treating phrases as distinct entities rather than collections of constituent words.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If conventional machine translation methods are used to translate text from one language to another, then translation speed and simplicity are improved, but accuracy of meaning and contextual understanding deteriorate

Engineering Contradiction:
Improvetranslation speedVSAvoidaccuracy of meaning
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent segments the translation process into multiple stages: initial direct translation, context analysis, semantic similarity assessment, and alternative phrase generation. This segmentation allows the system to maintain fast initial translation while adding layers of accuracy improvement through contextual understanding and semantic analysis.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces an intermediary semantic analysis layer between the source text and target translation. This intermediary process analyzes context, tone, and semantic relationships to generate alternative phrases that better convey the intended meaning, acting as a mediator that bridges the gap between speed and accuracy.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Device complexity

If word-by-word translation is used, then translation simplicity is improved, but loss of contextual information increases

Engineering Contradiction:
Improvetranslation process simplicityVSAvoidcontextual information loss
Core Design Contradiction:
Device complexityVSLoss of information

Solution Approach 1:

The patent applies preliminary context analysis before finalizing the translation. By analyzing the contextual information, semantic relationships, and tone of the source text before generating the translation, the system preserves contextual meaning while maintaining a relatively simple translation process.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent changes the parameter of translation from simple word substitution to semantic equivalence based on context. By adjusting the translation parameter to consider semantic similarity and contextual factors, the system reduces information loss while keeping the process manageable through automated analysis.

Inventive Principle:
Principle #35Parameter changes

3Speed

If direct translation of phrases is used, then translation speed is improved, but semantic accuracy deteriorates

Engineering Contradiction:
Improvetranslation speedVSAvoidsemantic accuracy
Core Design Contradiction:
SpeedVSReliability

Solution Approach 1:

The patent implements a feedback mechanism where the system generates initial translations, analyzes their semantic accuracy against the source text context, and automatically generates alternative phrases when semantic equivalence is insufficient. This feedback loop improves semantic accuracy while maintaining speed through automated processes.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent makes the translation process dynamic by adjusting the level of analysis based on detected user difficulty. When users struggle with translated phrases, the system dynamically generates alternative explanations and semantic equivalents, adapting the translation depth to maintain both speed and accuracy.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentUS10936635B2Context-based generation of semantically-similar phrases
Publication Date: 2021.03.02 INTERNATIONAL BUSINESS MACHINE CORPORATION
  • US10936635B2 patent drawing
  • US10936635B2 patent drawing
  • US10936635B2 patent drawing

AI summary

A computer system processes a document. A user viewing a displayed text is monitored and it is determined that the user is experiencing difficulty understanding a query phrase in a displayed text, wherein the displayed text is translated from a first language to a second language. One or more alternative phrases are generated for the query phrase by determining a plurality of candidate phrases based on the context of the query phrase and a tone of the displayed text, scoring each of the plurality of candidate phrases, and selecting a candidate phrase to be the alternative phrase based on the score of the candidate phrase. The one or more alternative phrases are presented to the user. Embodiments of the present invention further include a method and program product for processing a document in substantially the same manner described above.