MemorySphere, Translation Memory integrated with Machine Translation, AppTek.com

Products

MemorySphere

memoryspherepg

Integrated Translation Memory Software

Translators rarely translate documents that are completely new and unrelated to any previously seen texts. For frequently appearing phrases, such as idioms and domain-specific expressions, an ideal aid to the translator is a way to re-use previous translations.

AppTek Translation Memory provides a human-quality dimension to machine translation. The AppTek Translation Memory system scans a source text to match sentences and phrases against a database of pre-translated text, then presents relevant translations along with the location of the source material.

Components

AppTek Translation Memory is comprised of three components. The first component is the dynamic lexicons that are used to align and index translated text. The second component is the main database, which is used to store pre-translated text. The database can support multiple language pairs and multiple domain classifications for pre-translated text.

The third component is the document management and user interface, which allow users to manage the source documents, add and edit pre-translated text at the sentence level, and maintain the lexicons and system parameters. An optional corpus of translation memory for several domains is also available. The system can also be integrated with the AppTek TranSphere® machine translation system.

Functions

AppTek Translation Memory is designed to deliver state-of-the-art functionality and to enhance translators’ productivity and throughput. The system has the ability to map entire phrases or sentences for specific contextual use. Furthermore, the system utilizes dynamic monolingual lexicons for automatic sentence alignment and displays source sentences aligned with the target translation. The AppTek Translation Memory system delivers the following capabilities:

Indexes complete or segmented sentences
Creates word and phrase matches using a bilingual Morphology Analyzer
Performs alignment and mapping on a word-by-word basis
Features auto alignment option
Allows for customization to each translator and editor style
Performs many-to-one and one-to-many word alignment( (i.e., with idioms and multi-word terms))
Eliminates noise words (e.g., “the,” “a,” “an”)
Includes intuitive user interface for fast and easy manual sentence alignment
Features automatic bilingual translation memory creation when adding sentences to translation memory
Uses exact match and partial or block match techniques
Multiple language pairs and multiple domains
Features document management functionality

Efficient Translation

Because the system can map entire phrases or sentences for specific contextual use, as a result, the same sentence never needs to be translated again. These translation productivity gains create direct savings in translation costs, regardless of whether the work is outsourced or done in-house. AppTek’s Translation Memory technology is integrated with MT introducing a human-quality dimension to machine output; savings in total translation costs are typically in the range of 30% to 50% and can be as high as 70%.

Users of AppTek Translation Memory achieve up to 50% increase in productivity, resulting in faster time-to-market, manage communication quality and consistency improvements in localized documentation through standardization of terminology used for translations.

Click Here for PDF