Intellego

From MozillaWiki
Jump to: navigation, search

Intellego is a machine translation project for the benefit of Mozilla and the Open Web.

Project details

Intellego is a machine translation (MT) initiative that seeks to unify existing open MT projects by providing a single platform for engine developers and a unified web service that hosts a number of different language pairs/engines/implementations in the back end. The Intellego platform will allow users to select from a number of open MT engines based on the most prominent MT methodologies in order to find the best target MT output for their on-the-fly translation. This will be accomplished by partnering with existing open MT projects and hooking into their infrastructures, freeing Mozilla of the requirement to develop and host MT engines and promoting the use of MT engines within the fragmented open ecosystem.

MT users are limited to using engines that follow a single MT methodology for all language pairs and content types. Studies have shown that a one-size-fits-all approach in MT does not provide the user with optimal translation output. Users need a single access point to different MT engines following different MT methodologies that will produce the best quality output by selecting the right engine for the right language pair. Intellego seeks to further establish an open MT ecosystem, as we feel it is the best way to quickly provide high-quality MT services to users on the web at the lowest cost and in a way that engages the open source community.

Intellego Platform

Intellego is a centralized machine translation platform that provides the following services:

  1. A unified web service for open source machine translation engines, allowing them front-end accessibility on the Web. This service will function as follows:
    • Users insert text into a text field and see it translated on the fly in an accompanying text field.
    • Users insert an URL into a text field, Intellego strips the text nodes from the DOM, runs the text nodes through Intellego, and returns a rendered page in a new tab containing target language text.
    • Users can select their preferred MT engine.
    • Intellego will intelligently determine the most appropriate MT engine given a language pair and content type.
    • Advanced mode allows users to select and arrange MT pipelines utilizing multiple MT engines
    • Terminology processing utility (from GSOC project) will ensure accurate and consistent terminology translation despite MT method.
  2. An easy and simple API for the following services
    • A widget for developers to add to their sites for automatic translation by Intellego MT engines
    • Allows mozilla l10n tools developers to link directly into hosted open MT services through Intellego
  3. Users can post-edit MT output in context and submit as feedback (leveraging Pontoon modules). Community can vote on suggested post-edits, and highest voted feedback is entered into the engine corpuses.

Explore the wiki for more details about the Intellego project's purpose and focus as well as the goals and milestones for the platform's development and the platform technical spec (in progress).

Resources

Team

These are the members of our Intellego team, with a brief overview of their relevant skills:

Jeff Beatty (gueroJeff) (gueroJeff)
Team lead.
Localization, organization, programming.
Majken Connor (Kensie) (Kensie)
Community outreach, evangelism.
Gordon P. Hemsley (GPHemsley) (GPHemsley)
Linguistics, programming, BCP 47 (language tags).
Mekki MacAulay (mekki) (mekki)
Strategic management, partnerships, grants, business collaboration, evangelism.

Project meetings

The Intellego team meets every week to discuss the progress of the project.

We also occasionally have sprint meetings, where we work on a particular aspect of the project for a long stretch of time.

For more information about meetings, see our meetings page.

Discussion