Artificial intelligence and crowdsourced translation platform
Disclosed herein is a translation platform making use of both machine translation and crowd sourced manual translation. Translation is performed on pages in an application. Manual translations are applied immediately to local versions of the client application and are either human reviewed or reverse machine translated and compared against the original text. Once verified, the translations are applied to all end-clients.
1. A method comprising:
extracting a passage of text in a first language from a first page;
tracking one or more mapped locations of text in the first page;
selecting one or more machine translation engine to which at least a portion of the passage of text is transmitted to translate text into a second language based on the passage of text;
submitting the passage of text to at least one machine translation engine;
receiving the machine translated text in the second language;
generating a second language page matching the first page wherein the passage of text is replaced by the machine translated text;
receiving at an administrator server user input including corrected text to at least a portion of the machine translated text;
updating at an the administrator server the machine translated text on the second language page based on the corrected text;
causing presentation of the second language page using the one or more mapped locations of the text to replace text in the first language with the text in the second language;
parsing the passage of text based on length of the passage of text; and
determining the machine translation engine to use for the passage of text from a plurality of machine translation engines based on the passage of text,
wherein determining the machine translation engine to use is based on inclusion of words from a predetermined list.
2. The method of claim 1 , wherein the first page comprises a message from one user to another, an email, or a webpage.
3. The method of claim 1 , wherein submitting the passage of text to the at least one machine translation engine comprises:
using a lexicon file in a machine translation of the passage of text.
4. The method of claim 3 , wherein using the lexicon file comprises:
submitting the lexicon file to the machine translation engine configured to affect generation of the machine translated text wherein a predetermined set of words contained within the passage of text are machine translated to corresponding terms indicated by the lexicon file.
5. The method of claim 1 , wherein extracting the passage of text in the first language from the first page comprises:
extracting from one or more resource files the passage of text.
6. The method of claim 1 , further comprises:
causing presentation of a first portion of a translation control in a user interface as a portion of a user interface; and
receiving the passage of the text in the first language.
7. The method of claim 6 , further comprises:
causing presentation of a second portion of the translation control as another portion of the user interface; and
receiving a list of results including one or more alternative machine translated text.
8. The method of claim 1 , further comprising:
generating the second language page as a second version of the first page in which the machine translated text is placed as a same location as corresponding text in the first page as a function of the mapped location of the corrected text.
9. The method of claim 1 , wherein parsing the passage of text further comprises:
parsing the passage of text based on inclusion of words from a predetermined list.