The document discusses the evolution of the translation technology landscape and how translation is becoming datafied. It notes that machine translation output is around 100 times larger than human translation and diversifying with new technologies. Translation production is growing faster in volume and the industry is growing in revenue but changing radically with new technologies and business models. Machine learning and data are playing a key role in improving automated translation and automation of the translation process. The document explores trends around algorithmic management of translation and datafied translation platforms.
3. Convergence of Markets: Translation Shifts Gears
21st Century Convergence
Luxury
Publisher-driven
translation industry
Mobile
Real-time
Personalized
Datafied Embedded
New payment models
Good enough
Continuous
+ 1,000 languages
Innovation Invaders
From 10,000 customers
who buy translation as
a ‘luxury’ product to 6
billion users who
consider translation
‘free’.
4. What is happening now
Drag and drop …
• Real-time customization of MT
Plug and play …
• Mobile apps for speech-to-speech
Translation as utility
• On every screen, in every app
Convergence of:
• Technologies: MT & S2S
• Markets: Consumer &
Business
• Business: Free and Premium
Many new start-ups
Insider and invader innovators
MT as an API: pay as you go or free for data
5. Translation production is growing faster (in words)
HT
Machine Translation output
(No human touch)
= 100 times bigger than
human
translation production *
* 30B times 100 = 3 Trillion words in translation volume.
Human Translation = TM + MT + PE
And diversifying with crowdsourcing, digital marketing, testing, va
transcreation, talent search
* 30B/$0.10 = 300 Billion words in translation volume.
Check: 300 Billion words / half million per translator = 600,000 translators
6. Industry is Growing (in $)
0
5
10
15
20
25
30
35
40
45
1980 1990 2000 2010 2020
… but changing radically
Translation PM QM Testing Digital Marketing
Crowdsourcing Talent Mgnt MT PE Transcreation
7. We want to publish a report:
TAUS Five Year Translation Innovation Roadmap
Report to be published in April, freely downloadable from TAUS web site.
Objectives:
1. Influence the ecosystem and promote innovation across the industry
2. Provide a reference for strategy planning for all stakeholders in the global translation industry
3. Setting priorities and providing ideas and guidance for TAUS industry-shared services
Objectives of the TAUS Industry Consultation
8. The TAUS Industry Consultation Process
1. Survey to collect ideas: we received 60 ideas from TAUS members and non-
members
Nov./Dec. 2016
2. TAUS categorized the ideas and grouped them under 13 themes or programs Dec. 2016
3. 200+ responses to survey to prioritize the programs. Jan. 10 2017
4. TAUS New Year’s Reception webinar: preview of outcome Jan. 17, 2017
4. TAUS Advisory Board selects 5 to 7 programs that will be further researched and
prepared
Jan. 18 2017
5. TAUS team plus industry experts research ideas and prepare briefing for the
Summit
Jan./Feb. 2017
6. Registration for the Summit needs to confirmed (50 participants) Jan. 31 2017
7. Participants will be engaged in some of the Summit preparations Feb./Mar. 2017
8. Industry Summit takes place in Amsterdam, chaired by an independent facilitator Mar. 22-24, 2017
9. TAUS Five Year Translation Innovation Roadmap report will be published. Apr. 30 2017
9. Results Survey Industry Consultation
8
97
90
4 3
Distribution of Responses
Academic Buyers Providers Gov-NGO Freelancers
All responses are from decision makers, director level and business owners. (Thanks to
very targeted and personal email invitations.)
10. TAUS Priorities
Top 5:
1.Machine Learning
2.Quality Dashboard & DQF
3.Machine Translation
4.Intelligent TM
5.Interoperability
Second rank:
6. Data Cloud
7. Translation and IT infrastructure
8. Open source tools
Out?
9. Academy
10. Driving adoption
11. Speech data?????
12. Regulations
13. Crowdsourcing
Comparing Priorities Board with Totals
Total
Board
1. 2.3.5.
4.6. 7.8.9. 10. 11.
13. 12.
11. Translation is datafied
Machines are learning
For improvement of automated translation: Neural MT of course ….
… but also for translation process automation:
• Translator: productivity (words and edits per hour, mouse clicks), behavior, on-
time delivery, eye tracking, biographical data, capacity, third party evaluations,
interests, social graph
• Environment: weather, time zone, news events
• Project metadata: industry, content type, language pair, process, technology,
client, translator, reviewer
• Quality: requested quality level (good enough, human quality), purpose, source
analysis
• Technology: MT engines, TM leverage, process settings, data used for training
Lights out:
Translation becomes ‘invisible’, a utility like electricity, internet.
We (as professionals) will all have dashboards to control and track
the quality of our translations.
14. Discussion
Trends and innovations emerging from algorithmic management in the
translation industry?
Define Machine Learning: …..
Algorithmic management: “how to instruct, track and evaluate a
crowd of casual workers you do not employ, so they deliver a
responsive, seamless standardized service.”
15. SWOT of Machine Learning in Translation
S W
TO
• xxx
• xxx
• xxx
• xxx
16. This slide may not be used or copied without permission from TAUS