Machine Translation Reliability Feedback Loop

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Machine translation methods often produce errors that are difficult for users to understand, as they lack effective mechanisms for determining the reliability of translation results and re-determining accurate sentences.

Innovation Solution

A processor-implemented machine translation method that translates a source sentence into a target language, assesses the reliability of the translation, and re-determines the target sentence by retrieving similar sentences from a database, adjusting search techniques, or using a third language, and outputs the re-determined sentence as text or speech.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If machine translation is performed using conventional methods, then translation speed is maintained, but translation accuracy and reliability deteriorate due to errors that are difficult for users to understand

Engineering Contradiction:
Improvetranslation reliabilityVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent implements a feedback mechanism where the translation result is evaluated for reliability, and if unreliable, the system automatically re-determines the translation using alternative methods such as retrieving similar sentences from a database or adjusting search techniques, creating a closed-loop quality control system

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The system performs preliminary actions by pre-storing similar sentences and their translations in a database, and pre-defining multiple search techniques and translation models, so that when reliability is questioned, re-determination can quickly utilize these pre-prepared resources without significant additional complexity

Inventive Principle:
Principle #10Preliminary action

2Manufacturing precision

If the translation system re-determines inappropriate translations using multiple methods, then translation accuracy improves, but processing time increases

Engineering Contradiction:
Improvetranslation precisionVSAvoidprocessing time
Core Design Contradiction:
Manufacturing precisionVSLoss of time

Solution Approach 1:

The system applies partial re-determination by first attempting the quickest correction method (retrieving similar sentences from database) and only proceeding to more time-consuming methods (adjusting search techniques or using third language) if necessary, thus achieving sufficient precision without always incurring maximum time cost

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The re-determination process is segmented into multiple independent methods (database retrieval, search technique adjustment, third language translation), allowing the system to select and apply only the necessary segments based on the specific translation problem, reducing overall processing time

Inventive Principle:
Principle #1Segmentation

3Ease of operation

If the system uses database retrieval and multiple re-determination methods, then user understanding of translation results improves, but device complexity increases

Engineering Contradiction:
Improveuser understandingVSAvoidsystem complexity
Core Design Contradiction:
Ease of operationVSDevice complexity

Solution Approach 1:

The patent introduces an intermediary reliability evaluation component that acts as a mediator between the translation engine and the user, assessing translation quality and automatically triggering re-determination when needed, thereby improving user understanding without requiring users to directly interact with complex re-determination mechanisms

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentUS11093719B2Machine translation method and apparatus
Publication Date: 2021.08.17 SAMSUNG ELECTRONICS CO LTD
  • US11093719B2 patent drawing
  • US11093719B2 patent drawing
  • US11093719B2 patent drawing

AI summary

A processor-implemented machine translation method includes translating a source sentence expressed in a source language into a target language to determine a target sentence, determining, based on a reliability of the target sentence, whether the target sentence is appropriate as a translation result of the source sentence, and re-determining a target sentence corresponding to the source sentence in response to a result of the determining being that the target sentence is inappropriate as the translation result.