Your SlideShare is downloading. ×

TAUS OPEN SOURCE MACHINE TRANSLATION SHOWCASE, Beijing, Yu Gong, Adobe, 23 April 2012

1,175

Published on

Moses Tool Set is a set of tools to simplify the usage of Moses. By using this tool, the training process of Moses can be done in an easier and intuitive way. It consists of 4 features: Corpus Clean …

Moses Tool Set is a set of tools to simplify the usage of Moses. By using this tool, the training process of Moses can be done in an easier and intuitive way. It consists of 4 features: Corpus Clean Tool, Corpus Splitting Tool, Moses Training Harness, and Moses Scoring Harness. Each feature cannot only work independently but be combined into a job, which enables users to complete the whole training process in one click.

This presentation is a part of the MosesCore project that encourages the development and usage of open source machine translation tools, notably the Moses statistical MT toolkit.
MosesCore is supporetd by the European Commission Grant Number 288487 under the 7th Framework Programme.
Latest news on Twitter - #MosesCore

Published in: Technology, Art & Photos
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total Views
1,175
On Slideshare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
26
Comments
0
Likes
0
Embeds 0
No embeds

Report content
Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
No notes for slide

Transcript

  • 1. Moses Tool Set A set of tools based on Adobe technology to simplify your usage of Moses Yu Gong | Software Engineer© 2012 Adobe Systems Incorporated. All Rights Reserved. Adobe Confidential.
  • 2. Agenda §  Addressing Moses Pain Points §  Advantages of Moses Tool Set §  Moses Tool Set Architecture §  Moses Tool Set Features §  Useful Resources §  Q&A© 2012 Adobe Systems Incorporated. All Rights Reserved. Adobe Confidential.
  • 3. Addressing Moses Pain Points 1.  Corpus Cleaning 2.  Engine Training 3.  Engine Testing 4.  Integrating Moses With Linguistic Platform© 2012 Adobe Systems Incorporated. All Rights Reserved. Adobe Confidential.
  • 4. Advantages of Moses Tool Set •  User Friendly •  Platform Independent •  Open Source© 2012 Adobe Systems Incorporated. All Rights Reserved. Adobe Confidential.
  • 5. Moses Tool Set Architecture© 2012 Adobe Systems Incorporated. All Rights Reserved. Adobe Confidential.
  • 6. Moses Tool Set Features – Corpus Cleaning Moses  Func*onality   •  Tokenizing   •  Casing     •  Long  Segments   Adobe  Func*onality   •  Placeholder  Handling   •  URL  Handling   •  Number  Cleaning   •  Duplicate  Line   Cleaning   •  Weird  Aligned  Pairs   •  Cleaning  by  regular   expressions  © 2012 Adobe Systems Incorporated. All Rights Reserved. Adobe Confidential.
  • 7. Moses Tool Set Features – Corpus Splitting & Uploading Split  Corpus  by  Purpose   •  Training     •  Tuning   •  TesCng   Upload  Split  Corpus  to   Moses  Server  © 2012 Adobe Systems Incorporated. All Rights Reserved. Adobe Confidential.
  • 8. Moses Tool Set Features –Training & Tuning Command Line Pain Human Unfriendly •  Highly Detailed •  Error Prone •  Difficult To Reproduce© 2012 Adobe Systems Incorporated. All Rights Reserved. Adobe Confidential.
  • 9. Moses Tool Set Features –Training & Tuning UI  To  Simplify  Inputs   •  Training  Run  ID   •  Language  Model   Parameters   •  Corpus  ID   •  Source  &  Target   •  Default  Alignment   •  Default  Reordering   •  Remote  Server  © 2012 Adobe Systems Incorporated. All Rights Reserved. Adobe Confidential.
  • 10. Moses Tool Set Features –Testing •  How do you know when an engine is good enough? •  How do you know when it is intrinsically flawed? •  How do you automate comparing a new engine to old ones?© 2012 Adobe Systems Incorporated. All Rights Reserved. Adobe Confidential.
  • 11. Moses Tool Set Features –Testing   •  Reliable  Scoring   •  Bleu/Nist/Meteor   •  Simplified  UI   •  Dynamic  ConnecCon  to   exisCng  engines   •  Repeatable  © 2012 Adobe Systems Incorporated. All Rights Reserved. Adobe Confidential.
  • 12. Moses Tool Set Features –Testing   •  Reliable  Scoring   •  Bleu/Nist/Meteor   •  Simplified  UI   •  Dynamic  ConnecCon  to   exisCng  engines   •  Repeatable  © 2012 Adobe Systems Incorporated. All Rights Reserved. Adobe Confidential.
  • 13. Automation Corpus Cleaning Corpus Splitting & Uploading Training & Tuning Testing© 2012 Adobe Systems Incorporated. All Rights Reserved. Adobe Confidential.
  • 14. Localization Workflow Integration Moses Tooling Chain Linguistic Platform© 2012 Adobe Systems Incorporated. All Rights Reserved. Adobe Confidential.
  • 15. Resources Source Code: http://code.google.com/p/m4loc© 2012 Adobe Systems Incorporated. All Rights Reserved. Adobe Confidential.
  • 16. Questions© 2012 Adobe Systems Incorporated. All Rights Reserved. Adobe Confidential.
  • 17. © 2012 Adobe Systems Incorporated. All Rights Reserved. Adobe Confidential.

×