This presentation was given by Carlos Turró Ribalta, Head of Media Services at Universitat Politecnica de Valencia, Spain on 11 December at the REC:all workshop 2013 "Lecture Capture: Moving beyond the pilot stage: large-scale implementation of lecture capture in European Higher Education" in Leuven, Belgium.

1. Rec: All Lecture Capture Workshop
11 December 2013
Carlos Turró
Universitat Politècnica de València
EC FP7 ICT project #287755

2. Motivation
• Video lecture repositories and MOOCs
• Thousands of hours of video lectures available
• Hundreds of hours of video lectures recorded
every week
• Most video lectures only available in their
original language
• No subtitles
12 Nov 2013
2

3. Motivation
• Transcriptions and translations are needed
•
•
•
•
•
Accessibility for people with disabilities
Accessibility for speakers of different languages
Search and analysis functions
Automated topic finding
…
12 Nov 2013
3

4. Motivation
• Transcriptions and translations are needed
•
•
•
•
•
Accessibility for people with disabilities
Accessibility for speakers of different languages
Search and analysis functions
Automated topic finding
…
• How do we get there?
12 Nov 2013
4

5. The transLectures approach
1. Automatic Speech Recognition (ASR)
and Machine Translation (MT)
• Adaptation: Taking advantage of the characteristics
of video lecture repositories
• High-quality automatic transcriptions and translations
2. Interactive postediting:
intelligent interaction for reduced effort
12 Nov 2013
5

6. Goals
• Development of an engine for adaptation &
Intelligent interaction
• Implementation
• Case studies: Videolectures.NET & Polimedia
• Real-life evaluation
• Integration into Opencast Matterhorn
http://opencast.org/matterhorn/
12 Nov 2013
6

7. The transLectures partners
Name
1
2
3
3+
4
5
6
Country
Universitat Politècnica de València
Xerox SAS
Institut Jožef Stefan
Knowledge for All Foundation
RWTH Aachen University
EML – European Media Laboratory
DDS – Deluxe Digital Studios
Spain
France
Slovenia
UK
Germany
Germany
UK
Now we are in M25
12 Nov 2013
7
36 Months

8. Statistical Transcription
(and translation)
Acustic
Model
Sound
Language
Model
ASR Engine

9. Statistical transcription
(and translation)
Acustic
Model
Manually transcripted
voice
Modeling Engine
Language
Model

10. Architecture of TransLectures
Lecture
Language
Model
Result
Slides
Transcription
Translation
Extra
content
Intelligent interaction

11. Languages
• Transcription (ASR)
• EN
• SL
• ES
• Translation (MT)
•
•
•
•
EN>SL , SL>EN
EN>ES , ES>EN
EN>FR
EN>DE
12 Nov 2013
1
1

12. Case study:
VideoLectures.NET
15000 lectures

13. Case study: Polimedia
10000 Learning Objects

14. Demo
http://translectures.videolectures.net
http://polimedia.upv.es/catalogo
http://translectures.eu/player/

15. Scientific evaluations
• Transcription results
Worse
• WER: Word Error Rate (%)
• Goal: WER < 20%
• EN, SL, ES
Better
12 Nov 2013
15

16. Scientific evaluations
• Translation results
Better
• BLEU
• Goal: BLEU > 30
•
•
•
•
EN>SL , SL>EN
EN>ES , ES>EN
EN>FR
EN>DE
Worse
12 Nov 2013
16

17. Y2 results and comparison
12 Nov 2013
17

18. Y2 results and comparison
12 Nov 2013
18

19. Y2 results and comparison
12 Nov 2013
19

20. Massive adaptation
• Characteristics
of video lectures
Just one person
Known speaker
Clear talking
No interruptions
Focused on a topic
Slides
12 Nov 2013
20

21. Massive adaptation
• Known speaker and topic
• Slides
• Related documents
12 Nov 2013
21

22. Intelligent interaction
• Postediting automatic transcriptions/translations
• The user invests the least possible effort
• The system learns the most from it
• Confidence measures
• Fast constrained search
12 Nov 2013
22

23. Intelligent interaction
12 Nov 2013
23

24. Intelligent interaction
12 Nov 2013
20

25. Implementation and integration
• Videolectures.NET
• Polimedia
• Opencast Matterhorn
12 Nov 2013
25

26. The tL player
Online HTML5 VideoPlayer editor with editing capabilities.
The user interface has three different editing layouts, and full
keyboard support.
User interaction statistics analyzed to improve user
experience and develop a user model.

27. tL player

28. Manual upload of lectures

29. transLectures: tools available
• The transLectures-UPV Toolkit (TLK) for ASR
• www.translectures.eu/tlk
• RWTH Aachen: rASR, Jane (MT)
• http://www-i6.informatik.rwth-aachen.de/web/Software/
Note that you need an acoustic & language model
12 Nov 2013
29

30. transLectures: tools at M30
• The tL player (& editor)
• tL Opencast Matterhorn module
• Cloud service for testing
• Coming soon at M30 (www.translectures.eu)
More info at the
OCWC
conference
(Ljubljana) in
April 2014

31. Next steps for transLectures
• Keep improving ASR and MT results
• Keep improving tL open source tools (TLK, tL player)
• External user evaluations (VL.NET and polimedia)
• External trials: implementation in other universities
12 Nov 2013
31

32. Next EU project: EMMA
• MOOC related project
• transLectures work in adding 7 new transciption systems
(English, Italian, Spanish, French, Dutch, Portuguese and
Estonian)
• … and 8 translation systems (from
Italian, Spanish, French, Dutch, Portuguese and Estonian into
English; and from English into Italian and Spanish)
• Beginning in 2014

33. www.translectures.eu
Thanks!
My mail (Carlos Turro)
Project coordinator:
Alfons Juan-Ciscar
turro@cc.upv.es
ajuan@dsic.upv.es
EC FP7 ICT Programme – Project Number 287755
12 Nov 2013
33