Language Technology: Democracy Participation Collaboration. Talk by Gareth Morlais at Idiap Research Institute 19/01/2016
Digital storytelling, Welsh Government's Welsh-language Technology and Digital Media action plan
The Codex of Business Writing Software for Real-World Solutions 2.pptx
Language Technology: Democracy Participation Collaboration. Talk by Gareth Morlais at Idiap Research Institute 19/01/2016
1. Corporate slide master
With guidelines for corporate presentationsWelsh PPT template
The title slide of your Welsh language PowerPoint presentation should contain the Welsh
Government logo and Welsh URL address as positioned here on the red template areas.
Do not alter the size or position of these areas.
You are NOT REQUIRED to put th branding on subsequent slides in your presentation
Welsh Government
Democracy
Participation
Collaboration
Gareth Morlais at
IDIAP 19/01/2016
12. DIGITAL STORIES
•True personal stories
•Told by the individual it happend to
•About 250 words, 2 mins., c.20 photos
•Change from paternalistic model of
broadcaster telling stories on people’s
behalf
•Refreshing, surprising.
16. Cynllun Gweithredu Technoleg a
Chyfryngau Digidol Cymraeg
Welsh-language Technology and
Digital Media Action Plan
Marchnata a chodi
ymwybyddiaeth
Marketing and awareness raising
Ysgogi’r prif gwmnïau technoleg Motivating the main technology
companies
Ysgogi datblygiad o becynnau
meddalwedd a gwasanaethau
digidol Cymraeg newydd
Encouraging the development of
new Welsh-language software
applications and digital services
Ysgogi creu, rhannu a defnyddio
cynnwys digidol Cymraeg
Stimulating the creation, sharing
and consumption of Welsh-
language digital content
Cefnogi arfer da yn y sector
cyhoeddus, y sector preifat a’r
trydydd sector
Supporting good practice in the
public, private and third sectors
25. Sut i oroesi yn yr oes ddigidol. Canllaw i ieithoedd
bit.ly/digiskit
26. DIGITAL SURVIVAL KIT FOR YOUR LANGUAGE
E.g.
Dictionaries, term dictionaries, corpora, encyclopedias such as Wikipedia
Speech command recognition
Speech-to-text
Text-to-speech / Synthetic voice
Automatic semantic analysis and tagging
Artificial Intelligence
and AI-to-speech linkage systems
Machine translation
Social media aggregation (like ffrwti.com)
Local publishing platforms (like poblcaerdydd.com)
Data about geographical linguistic resources
Interactive embeddable maps
etc.
27.
28.
29.
30. £2million project
10 million words
Transcriptions of casual spoken and written Welsh
New analyses used to steer language learning
32. SOLUTION?
•Look to reuse openly-licensed resources
•Use or build enablers; separate content
and structure
•Seek collaboration from research
institutes &universities
•Invite feedback and partnership
so others can tailor their own
•See if it will scale up
•Industrialize the process (e.g. word-a-day)
55. The purpose of getting power
is to be able to give it away.
- Aneurin Bevan
56. Corporate slide master
With guidelines for corporate presentationsWelsh PPT template
The title slide of your Welsh language PowerPoint presentation should contain the Welsh
Government logo and Welsh URL address as positioned here on the red template areas.
Do not alter the size or position of these areas.
You are NOT REQUIRED to put th branding on subsequent slides in your presentation
Welsh Government
Democracy
Participation
Collaboration
Gareth Morlais at
IDIAP 19/01/2016
Diolch
Merci
Merci vielmal
Dangge
Danke
Thanks
Editor's Notes
1
Welsh Government’s Welsh-language Technology and Digital Media Action Plan has been
5 Amcan Cynllun Gweithredu Technoleg a Chyfryngau Digidol CymraegThe 5 Aims of the Welsh Government’s Welsh-language Technology and Digital Media Action Plan
These further education students at Coleg Meirion Dwyfor in Pwllheli, north-west Wales, are learning how to build their own Welsh Android apps using MIT’s App Inventor. They are then going into local primary schools to demonstrate what they built to young children and they show the steps they took to build their app.
This really worries us. Digital language extinction is a concern for our languages. We’ve found Meta-Net’s White Paper series useful in helping to identify gaps and opportunities. The Welsh edition’s author is Dr Jeremy Evas of Cardiff University.
If there’s a threat of extinction, you need a survival kit...
Here are some of the tools in the survival kit. Some of the text resources are widely found in many languages; the speech technologies such as speech to text are more expensive to commission.
Wales’s First Minister Rt. Hon Carwyn Jones AM (centre) with Dewi Bryn and Delyth Prys of Bangor University’s Language Technology Unit launching Paldaruo, an app which crowdsources audio recordings of Welsh sentences. It’s being used to train Bangor’s speech recognition resource.
This £30 Maplin robot arm executes Welsh voice commands via this Raspberry PI, thanks to Bangor University’s Language Technology Unit’s Gallu project, funded in part by Welsh Government’s Welsh-language Technology and Digital Media Grant.
Transifex was the crowdsourcing community translation tool we used to do this.
By building the infrastructure so it can be rolled out in multiple languages and then speakers of each language can go ahead and populate it with content. This is an important way forward as everyone’s resources become more limited.
What if we could crowdsource speech recognition in the same way as Transifex deals with text translation. After professional lanuage technologists have segmented and roughly-aligned the text and audio, let’s build platforms which ask people at home to refine the alignment, correct errors. This could be turned into a game.