Moses Machine Translation is finding its way to the localisation industry, thanks to its promise of offering a cheap (Open Source) and fast (statistical-based) solution. With the numerous Internet sources that walk novice users through their first steps, setting up and training a Moses engine has become increasingly straightforward. However, when it comes to deploying Moses in a live production environment, there is very little documentation to be found. The available tutorials are restricted to a prototyping environment, and tend to get too theoretical and technical for non-expert users.
Professional translation agencies who want to take full control over their Machine Translation Strategy based on Open Source technology need practical, hands-on information. Therefore, for the fourth year in a row, CrossLang organises a comprehensive training course "build your own Machine Translation system".
Venue
Vlerick Management School, Reep 1, 9000 Ghent - Belgium
24 March 2014
9:00 AM - 5:30 PM
Cost
The cost to participate in this full-day training course is 950 euro and includes lunch, coffee and refreshments, plus a syllabus and/or hand-outs.
Target audience
Localisation (tools) engineers, technical programme managers, computational linguists, CAT specialists with a technical background
Be sure to sign up now before we run out of space. If you want to be the first to know about our next training events leave your email address.
Alternatively, contact Nathalie De Sutter (+32 486 515113 - nathalie.de.sutter@crosslang.com) for more information on our dedicated, in-company training programmes.
About CrossLang Training Centre
CrossLang has extensive experience in building and deploying Moses SMT systems and in training adopters of the technology. In our training sessions, we help you look past the current limitations of Moses SMT and demonstrate how you can harness the system to achieve the same service level provided by commercial systems. We will help you build a business case for the use of Moses SMT in your operations, provide hands-on experience using a live SMT system, explain the technical details and guide you on how to prepare the system for production use.
- Probability and Translation
- SMT as a Search Problem
- Word & Phrase Alignments
- Translation & Language Models
- Data Clean-up & Partitioning
- Training a Baseline SMT Engine
- Recasing, TrueCasing
- Minimum Error Rate Training (MERT)
- Reordering Models
- Tuning a Moses Engine (MERT)
- Training a Recaser
- Automatic Evaluation Metrics
- Generating and Evaluating MT Output
- Moses XML-Markup
- Domain Adaptation Strategies
