Microsoft Technology Licensing, LLC

Marcus Andrew Taylor

Enyuan Wu

Alan K Michael

Beom Seok Oh

Shusuke Uehara

Philip L Davidson

Daniel Doubrovkine

Kenneth Jordan

Dong Kwon Joo

Jürgen Eidt

Douglas Frankin

Li Tian

Neal Allen Stipe

Daniel Olewski

Stephen Faiks

Fernando De La Garza Martínez

Jason Abbatiello

Bhavishya Mittal

Prasidh Srikanth

John Bowser

A software input string is received and tokenized into a sequence of tokens. The sequence of tokens is applied to a trained sequence-dependent lock/unlock classifier so that each of the tokens is classified as a token that should be locked, or remain unlocked, for subsequent translation. The software input string is converted to a converted string, in which the locked tokens are identified and the converted string is submitted for machine translation. A machine translation result is received and converted so that the locked tokens are replaced in the machine translation result, to obtain a translated software string.

Machine translation locking using sequence-based lock/unlock classification

Techniques to enhance software production are described. An apparatus may include a processor and a memory unit to couple to the processor. The memory unit may store a software build service that when executed by the processor is operative to receive a request to subscribe to the software build service from a software development client, create a virtual software build platform to build a software application, and assign multiple build resources from a resource pool to the virtual software build platform to build the software application, with each of the multiple build resources having one or more assigned affinity relationships for build execution tasks to build the software application. Other embodiments are described and claimed.

Techniques to enhance software production

Input text may be broken into sentence, or other types of segments, by first detecting exceptions in the input text, and then detecting break positions. Given a segment breaking scheme that comprises a set of break rules and a set of exceptions, a regular expression is created that represents the break rules, and another regular expression is created that represents the exceptions. The input text is analyzed to identify strings that match any exception, and the matching strings are substituted with placeholders that are not likely to occur naturally in the input. The resulting text, with substitutions, is then evaluated to find the positions in the text that match the break rules. Those positions are declared to be segment breaks, and the placeholders are then replaced with the original strings. The result is the original text, with breaks assigned to the appropriate positions in the text.

Efficient use of exceptions in text segmentation

Growing community of inventors

Bonney Lake, WA, United States of America

Marcus Andrew Taylor