microsoft translator william lewis [email protected] kites symposium october 31, 2013 -...
TRANSCRIPT
2
Overview
Introduction to Microsoft Translator, Tools, Products, etc.
Extent of Localization - Methods of Applying MT
Collaborative MTAssessing QualityApplication in Knowledge BaseBuilding your own MTCollaboration with Language Communities
Why MT?The purpose
The Crude Extent of localization Data Mining & Business Intelligence Globalized NLP Triage for human translation
Research Machine Learning Statistical Linguistics Same-language translation
The Good Breaking down language barriers Text, Speech, Images & Video Language Preservation
NOT:
Spend less money Take the job of human
translators Perform miracles
Microsoft Translator – Quick Facts
Linguistically informed statistical MT system 41 languages – from any language to any other language Runs in Microsoft Datacenter Simple web service API: SOAP, REST, AJAX, OData, web site widget 2 million characters/month free Available in the Enterprise Agreement, as a monthly subscription For extreme confidentiality situations, available on-premise
Highly customizable:– Collaborative Translations – Involve community, coworkers and customers– Hub: Custom engine training via an easy-to use UI
Web Scale– Powers translations in Bing, Microsoft Office, Microsoft SharePoint, Internet
Explorer, Yammer– Powers translations in Facebook, Twitter, eBay, and many other government and
enterprise sites
4
Microsoft Translator at a Glance
World-class Statistical Machine TranslationBuilt on over a decade of work at Microsoft Research
Big Data PoweredTrained with billions of “parallel” sentences (Bing index & licensed)
General Purpose SystemPowers Bing Translator, supports 40+ languages, any-to-any
Unprecedented Customization CapabilityHub (train before translation) + CTF (edit after translation)
Powerful Cloud APIRich, secure API enabling integrations, 99.9% availability
Fully integrated across the stack, Translator extends the value of Microsoft platform and your solutions built on the Microsoft platform for our customers including consumer facing applications such as Bing Translator, Bing Toolbar, Bing Dictionary, and Windows Phone App.
+80,000 more.
A few of our customers and partners….
Enabling Translation in Many Products
Powerful Tools and Customization
Our machine learning & big-data based translation technology brings the power of instant translations to break down language barriers for users, developers, webmasters, translators and businesses. Robust, industry leading tools such as the HUB and CTF allow for unprecedented customization of the translation experience.
Translation
Instant translation and language services in web, desktop and mobile applications.
Highly scalable and robust cloud-based, machine-translation service from Microsoft.
Supports SOAP, REST, AJAX, OData, and the Translator web site translation widget.
Extensibility for development on SharePoint, Office , Windows Phone, and more…..
Instant translations of web pages without the need to write any code.
Use the AJAX API to roll-your-own widget.
Use the integrated “Collaborative Translations” (CTF) functionality to tap into your community.
Customization
Custom translation portal to build, train, and deploy customized automatic language translation systems.
Combine your data with Bing big data to tune the translation output to best fit your content.
Free with any level of Translator subscription (including the free tier).
Override, modify or vote for the translated output to best fit the content.
Provide the end-user alternative translations.
Import the edits back into Hub for further training.
Hub CTFWidgetPowerful API
Integrates with your TM tool
8
Top translation tools support Microsoft Translator
Give these a try! (Demo)
Bing Translator for Windows PhoneThe Most Innovative Translation App on any Phone
Lync Conversation TranslatorRealtime Multi-lingual Conversations with Lync
Translator Widget for WebpagesInstant On-demand Translation for any Web site
Word Web App (Microsoft Office)Rich Web based Document Translation, now available in SharePoint, Outlook.com & SkyDrive
Contextual ThesaurusUtilize the Power of Machine Translation to Translate “English to English”
10
PriceCompetitively priced
Monthly subscription Free for up to 2 million characters per month Base price: $10 per million characters Discounted for higher volumes Paid by credit card or via Microsoft Enterprise agreement
11
Extent of localizationMethods of applying MT
Post-Editing
Goal: Human translation quality
Increase human translator’s productivity
In practice: 0% to 25% productivity increase– Varies by content, style and
language
Raw publishing
Goals:– Good enough for the
purpose– Speed– Cost
Publish the output of the MT system directly to end user
Best with bilingual UIGood results with technical audiences
12
Extent of localizationMethods of applying MT
Post-Editing
Goal: Human translation quality
Increase human translator’s productivity
In practice: 0% to 25% productivity increase– Varies by content, style and
language
Raw publishing
Goals:– Good enough for the
purpose– Speed– Cost
Publish the output of the MT system directly to end user
Best with bilingual UIGood results with technical audiences
Post-Publish Post-Editing“P3”Know what you are human translating, and why
Make use of community– Domain experts– Enthusiasts– Employees– Professional translators
Best of both worlds– Fast– Better than raw– Always current
Always thereAlways currentAlways retaining human translationsAlways ready to take feedback and corrections----------Midori Tatsumi, Takako Aikawa, Kentaro Yamamoto, and Hitoshi IsaharaProceedings of Association for Machine Translation in the Americas (AMTA)November 2012
Collaboration: MT + Your community
Your community
Your Web Site
Your App
Microsoft Translator
Collaborative TM
Enormous language knowledge
Microsoft Translator API
TranslationRequest Response
Match first
Translate if no match
What makes this possible – fully integrated 100% matching TM
Collaborative TM entries: Rating 1 to 4: unapproved Rating 5 to10: Approved Rating -10 to -1: Rejected
1 to many is possible
Making it easier for the approver – Pending edits highlight
Making it easier for the approver – Managing authorized users
Making it easier for the approver – Bulk approvals
What is Important?In this order
QualityAccessCoverage
Measuring Quality: Human EvaluationsKnowledge powered by people
Absolute 3 to 5 independent human evaluators are asked to rank
translation quality for 200 sentences on a scale of 1 to 4– Comparing to human translated sentence– No source language knowledge required
24
4 Ideal Grammatically correct, all information included
3 Acceptable Not perfect, but definitely comprehensible, and with accurate transfer of all important information
2 Possibly Acceptable
May be interpretable given context/time, some information transferred accurately
1 Unacceptable Absolutely not comprehensible and/or little or not information transferred accuratelyAlso: Relative evals, against a competitor, or a previous version of
ourselves
Measuring Quality: BLEU*Cheap and effective – but be aware of the limits
A fully automated MT evaluation metric–Modified N-gram precision, comparing a test
sentence to reference sentencesStandard in the MT community
–Immediate, simple to administer–Correlates with human judgments
Automatic and cheap: runs daily and for every change
Not suitable for cross-engine or cross-language evaluations
25
* BLEU: BiLingual Evaluation Understudy
Result are always relative to the test set.
26
Measuring Quality In ContextReal-world data
Instrumentation to observe user’s behavior
A/B testingPolling
In-Context gives you the most useful results
27
Knowledge Base (since 2003)
29
30
Knowledge base feedback
Source: Martine Smets, Microsoft
Customer Support
31
Knowledge Base Resolve Rate
Human TranslationMachine Translation
Microsoft is using a customized version of Microsoft Translator
Statistical MT - The Simple View
TranslationEngine
Collect and store parallel and target
language data
Web mined data
Train statistical models
Government dataMicrosoft manuals
DictionariesPhrasebooks
Publisher data
High-Performance Computing Cluster
TranslationEngine
User InputText, web pages, Chat etc
Distributed Runtime
Translated Output
Translation APIs and UX
32
Collaboration: MT + Your community
Your community
Your Web Site
Your App
Microsoft Translator
Collaborative TM
Enormous language knowledge
Microsoft Translator API
Remember the collaborative TM? There is more.
Collaboration: You, your community, and Microsoft
Your community
Your Web Site
Your App
Your TMs
Your previously translated documents
Microsoft Translator
Collaborative TM
Microsoft Translator Hub
Enormous language knowledge
Microsoft Translator API
Your custom MT system
Your collaborators
You, your community and Microsoft working together to create the optimal MT system for your terminology and style
35
Multiple community models–Necessity: driven by crisis–Love of language: driven by strong language/cultural identification
–Preservation: desire to preserve language
Haitian CreoleWhite Hmong
Community-driven MT
One of two official languages in HaitiA creole that evolved from French, Spanish, and several African languages (large % French-like)
Spoken natively by most of Haiti’s 8M peopleRecent as a written language (first literature dates to late 18th century), growing literature base
Semi-literate population, with preference to French (until recently)
Somewhat inconsistent orthographyLimited (but growing) Web presence
Haitian Creole
The earthquake of January 12th, 2010 a significant humanitarian crisis.
Aid agencies, foreign governments, a variety of NGOs, all responded en masse
Tranbleman tè nan Pòtoprens, kapital Ayiti!
Moun ap fouye pami debri yon bilding ki kraze nan tranblemann' tè 12 Janvye a.
Pòtoprens te catastrophically afekte 12 janvye 2010 tranbleman tè a.
Need for translated materials critical, especially those related to medicine and the relief effort.
Mission 4636 text messages from the field (up to 5K/day at peak) require rapid translation
At 10:30 a.m. on Tuesday, January 19th 2010, our team received an e-mail from a Microsoft employee in the field:– Do we have a translator for Haitian Creole?– If not, could we make one?
A little soul searching:– No one on our team knew anything about Creole
• No native speakers• No linguistic background on the language• No idea about grammatical structure
– No idea about encoding or orthography– No knowledge about registers or the degree of literacy– No parallel or monolingual training data of any kind (nor readily
available documents we could start with)– In effect, we were starting at Zero
So what else could we do but say “YES!”
The E-mail
Emergency SMS infrastructure Setup immediately in wake of Jan.
12, 2010 quake
Mission 4636
Mission 4636:Received SMSsTranslatedCategorizedTriagedRouted to aid agencies
Fanmi mwen nan Kafou, 24 Cote Plage, 41A bezwen manje ak dlo
Moun kwense nan Sakre Kè nan Pòtoprens
Ti ekipman Lopital General genyen yo paka minm fè 24 è
Fanm gen tranche pou fè yon pitit nan Delmas 31
Mission 4636 Messages
My family in Carrefour, 24 Cote Plage, 41A needs food and water
People trapped in Sacred Heart Church, PauP
General Hospital has less than 24 hrs. supplies
Undergoing children delivery Delmas 31
Over 80,000 messages received, up to 5,000+/day
Crisis Infrastructure: Message Pipeline
SMS
Tweets
Media
MessagePortal
Crowd(Translate)
MT
Triage
Geolocate
Lewis et al, 2011
White Hmong: not a crisis scenario like Creole
But, a language in crisisSome background:
–The Hmong Languages–The Hmong Diaspora–Decline of White Hmong and its usage in younger Hmong
White Hmong
Involves two critical groups:–Community of native speakers–Community leader(s)
Wide spectrum of users across the Hmong community:–College students–High school students–School teachers–School administrators, deans, professors–Business professionals–Elders
Community Engagement
Locating and vetting data–Locate data–Review documents that contain Hmong data
–Review parallelism of Hmong-English documents
Actively correcting errors from the engineContributing translation “repairs” on web sites that translate to Hmong
Building MT: Community Contributions
Home page (Web page viewer, cut-and-paste translator)
Haitian Creole and Hmong are among the languages available through our API (Advanced Programming Interface)–Multiple interfaces: AJAX, SOAP, HTTP–Can integrate translation directly into a variety
of appsWidget
–Integrate translation into Web pages–Traffic kept client side
Tools Available for Haitian Creole and Hmong
Widget/Collaborative Translation Framework (CTF)–Community can contribute translations–These can be published to Web pages
–Mixes MT with “trusted” human translations
Tools Available for Haitian Creole and Hmong
52
Just visit http://hub.microsofttranslator.com to do it yourself
Contacts
Web sitewww.microsoft.com/translator
Licensing & Pricing [email protected]
General & Customer [email protected]
54