System and method for evaluating the quality of human translation through the use of a group of human reviewers
AssigneeOne Hour Translation, Ltd.
InventorOfer Shoshan
Life of this patent
Estimating the term…
Abstract
System and computer implemented method for evaluating the quality and accuracy of a human translation from a source language to a target language by a group of professional translation reviewers communicating in real-time. The system and method comprise a system computer with a database of profiles and credentials of reviewers; and the ability to communicate with groups of selected reviewers via a computer network. The computer system assigns evaluations and collects ratings returned by reviewers, and calculates a consensus rating for the group, while omitting statistical outliers. The system will also calculate a score for the reviewer as compared to the group, and use the score in selecting reviewers for future evaluations.
Claims19
A networked based computing system for a group of professional translator reviewers to evaluate a quality of a language translation of text conducted by a human, comprising:
- a) a system computer comprising a processor and a memory device operably connected to one another, and a plurality of computer-executable instructions stored on the memory device and executed by the processor, comprising:
- i) a rating collection module that collects a rating on the quality of said language translation from each of the professional translator reviewers within said group;
- ii) a rating calculation module that calculates a consensus rating of the professional translator reviewers' ratings;
- iii) a presentation and action module that sends the consensus rating to a client computing system and takes corrective action when the consensus rating is below a quality threshold;
- and,
- iv) a reviewer scoring module that calculates a score for each of said professional translator reviewers;
- and,
- v) a database storing said professional translator reviewers' profiles and credentials;
- b) a plurality of computing devices in electronic communication with said system computer;
- and
- c) a client computing system in communication with said system computer to send a request for an evaluation, and to receive said consensus rating on a designated translation project.
A computer implemented method for a group of professional translator reviewers to evaluate a quality of a language translation of text conducted by a human, comprising:
- a) transmitting a translation evaluation request from a client computing device to a system computer;
- b) under control of a system computer database, identifying a plurality of professional translator reviewers with qualifying credentials;
- c) communicating with said plurality of the professional translator reviewers via wired or wireless electronic communications;
- d) evaluating the quality of the translation by the plurality of professional translator reviewers and each reviewer assigning a quality rating;
- e) calculating a consensus rating by the system computer for a collection of quality ratings from said plurality of professional translator reviewers;
- f) calculating a reviewer score by the system computer based on a comparison of each professional translator reviewer's rating to the consensus rating;
- and,
- g) forwarding the consensus rating to said client computing device.
A computer system comprising a processor and a memory device operably connected to one another, and a plurality of computer-executable instructions stored on said memory device and executed by said processor, comprising:
- a) a rating collection module that collects a rating on a quality of a human translation of text from each member of a group of professional translator reviewers;
- b) a rating calculation module that calculates a consensus rating of said group's ratings;
- c) a presentation and action module that sends the consensus rating to a client computing system and automatically orders a new translation when the consensus rating is below a quality threshold;
- and,
- d) a reviewer scoring module that calculates a score for each reviewer within said group based on a comparison of each reviewer's quality rating and the group's consensus rating.
Dependencies are inferred from the claim text.