Easing transcripts for mooc videos with an asr lwmoo cs
-
Upload
carlos-turro-ribalta -
Category
Engineering
-
view
257 -
download
0
Transcript of Easing transcripts for mooc videos with an asr lwmoo cs
![Page 1: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/1.jpg)
Easing Transcripts for MOOC Videos with an ASR (Automated
Speech Recognition) System
Carlos Turró, Jorge Civera and Jaime BusquetsUniversitat Politècnica de València
![Page 2: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/2.jpg)
![Page 3: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/3.jpg)
![Page 4: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/4.jpg)
![Page 5: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/5.jpg)
![Page 6: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/6.jpg)
The result of not having a screwdriver
• Pain• Frustration• Select a different tool
![Page 7: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/7.jpg)
How can I transcribe a video?• Manually transcribing a video
takes 10 times the length of the video (RTF)
• Boring• It’s worse if you don’t know
about the topic of the video
![Page 8: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/8.jpg)
Automated Speech Recognition (ASR)• How good is it?• Will it recognize my special
words? • Will it really help me?
![Page 9: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/9.jpg)
UPValenciaX MOOCs - Transcribing
https://media.upv.es/?id=b444d12e-db23-9a4f-9b3b-d1d9275d4cb4
![Page 10: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/10.jpg)
UPValenciaX MOOCs - Transcribing
https://www.youtube.com/watch?v=dKrbzX5NjTs
![Page 11: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/11.jpg)
UPValenciaX MOOCs - Transcribing
30 MOOC courses
![Page 12: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/12.jpg)
UPValenciaX MOOCs -Transcribing
• API• Just after
recordingASR
• RTF 3• Teaching
AssistantsReview
![Page 13: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/13.jpg)
UPValenciaX MOOCs –Transcribing
• API• Just after
recordingASR
• RTF 3• Teaching
AssistantsReview
70% less time
![Page 14: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/14.jpg)
Transcription and Translation Platform• Post-editing web interface (in HTML5)
![Page 15: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/15.jpg)
Crowdsourcing• We are crowdsourcing the on-campus courses using our own Paella
video player.
![Page 16: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/16.jpg)
How to get good transcription quality
•Transcription systems learn to transcribe from examples–At least 50 hours of videos (audio) in the source language previously transcribed
to learn the acoustic model–Texts in millions of words to learn the language model
Language Videos (hours) Text (Mwords)Dutch 532 628English 620 464000Estonian 130 410French 88 1800German 36 135Portuguese 54 573Italian 54 868Slovene 27 224Spanish 128 654
![Page 17: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/17.jpg)
How to get good transcription quality (II)
•Adaptation of transcription systems to the specific videos is key for high accuracy
•Availability of videos manually transcribed with similar acoustic conditions•Availability of text resources related to the video in question
· Title is used to retrieve related documents· Slides contain most of the special words used by the lecturer· Documents: text content from the course, additional text resources (bibliography)
• Sound quality of the video has a direct relationship with quality• No noise, no background music, please
![Page 19: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/19.jpg)
Our next step
Translations !!
![Page 20: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/20.jpg)
Conclusions• ASR technology is enough mature to help a lot in captioning• However, there should be a review phase• Quality can be enhanced by providing transcribed videos• At UP Valencia we got transcribed our 30 MOOC courses with 3x TA
cost
![Page 21: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/21.jpg)
Thanks!Questions?
![Page 22: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/22.jpg)
![Page 23: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/23.jpg)
Why transcription of MOOC video files?• Accessibility
![Page 24: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/24.jpg)
Why transcription of MOOC video files?• Accessibility
• Searching into a video file• Searching into a video repository• Topic identification• …and much more
![Page 25: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/25.jpg)
Measuring Quality: Word Error rate
WhereS is the number of word substitutions,D is the number of word deletions,I is the number of word insertions,N is the number of words in the reference text
![Page 26: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/26.jpg)
Measuring Quality: Word Error RateLanguage WEREnglishDutch
20.824.5
Italian 17.7Spanish 14.4Estonian 27.1French 22.7
![Page 27: Easing transcripts for mooc videos with an asr lwmoo cs](https://reader036.fdocuments.net/reader036/viewer/2022062820/58a423501a28abec1a8b656b/html5/thumbnails/27.jpg)
Attributions• Fingerspelling & tools Wikipedia• Bored https://www.flickr.com/photos/left-hand/3132070992/• Siri https://www.flickr.com/photos/smemon/8070397213/