Tony (Chief Architect, KantanMT.com) opens the proceedings with a temporal look at how MT technology has progressed. While embracing Rule Based MT in the 1970s, the industry switched over to Statistical MT around 2002 and is now faced with a new paradigm of Neural MT in 2016. For each technology progression, improved translation quality and fluency were achieved.
Summary: https://www.youtube.com/watch?v=19yyDa6mAsc
Full video: https://www.youtube.com/watch?v=EtbML0DTNHk
4. MT Platform
Cloud-based
Highly scalable for high volume
Fully Automated Useable Translation (FAUT)
Fusion of TM & MT & rules
Multiple Deployment Options
MT Factory for Rapid Development
Our Vision
To put Machine Translation
Customization
Improvement
Deployment
into your hands
What is KantanMT.com?
Active KantanMT Engines
10,394
Training Words Uploaded
534,179,753,587
Member Words Translated
8,571,269,925
www.kantanMT.com
13. The Fundamentals of Vectors
Neural Networks
Represent words as vectors (ie distinct numbers)
Distribution Analysis of words across corpus – context!
Semantically similar words will be mapped to nearby points
14. The Fundamentals of Vectors
Using a larger corpus
yields interesting results
University – College
Information – Data
Extending this to
sequences yields very
interesting results
This is at the core of
Neural MT
Dimitar will Deep-Dive
into this later
15. How do we know?
37
21
13
24
10
21
24
21
34
19
28
25.2
37
58
53
56
62
53.2
ENGLISH->CHINESE ENGLISH->JAPANESE ENGLISH->GERMAN ENGLISH->ITALIAN ENGLISH->SPANISH ALL
A/B Ranking SMT – NMT
Ranked by Professional Translators
Same SMT NMT