One of our largest projects, at the moment, is the open source-based speech synthesiser Wikispeech, which, e.g., will be used on Wikipedia. Wikispeech will be released in Swedish, English and Arabic. Dr. Nawar Halabi, at University of Southampton, has released a large part of the material needed, to make it possible to release it it in Arabic, under a free license. We decided to ask him a couple of questions.
Having Wikipedia with an Arabic Speech Synthesiser would encourage editors to correct and diacritise Arabic script on Wikipedia for it to be automatically read in a better way. This would make available more data for researchers to use in building their systems for Speech Synthesis and drive the technology forward.
– Nawar Halabi
– Nawar, what is your background, academic and personal?
– I was born in Aleppo, Syria 26 years ago. I lived there for 22 years of my life. I have went to school there and done my computer science degree. I also received two degrees in Classical Guitar performance. in 2012 I moved to Southampton, UK to start my MSc (Masters) in Web Technologies, then I started my PhD in 2013 and finished in September 2016. I worked as a research assistant on several projects along side my PhD.
– How come you started working on a speech synthesis project, and why do you think this kind of projects is important?
– I worked with the Assistive Technologies Group at the School of Electronics and Computer Science at the University of Southampton for 3 years. This has grown my interest in developing technologies to help empower and include people. This is not only for people with special needs but for everyone and speech synthesis was my way of doing this. Also, Speech Synthesis has a lot of interesting scientific knowledge around it. Machine Learning, Artificial Intelligence, Mathematics, Signal Processing etc. All of which are subjects of interest to me.
– My interest in machine intelligence started when I was young but I only found out about speech synthesis when I came to Southampton.
– For what reasons did you decide to make your work available under a free license?
– It is a struggle for every PhD student to find purpose for their work and to make other researchers find and reuse their work. Working with Wikimedia is certainly a way for me to spread the word about my work and leaving part of my work for free gives me the joy of feeling that I’m helping someone and encourages people to use my work.
– For those planning to release their work under a free license, what are the lessons you would like to give?
– DO IT.
– Having a widely used free tool increases your employablity and puts your name out there.
– Make sure you inform your sponsoring institution early on. As many institutions are very defensive of the work done at their premises.
– What are your expectations on the future of Wikispeech?
– I would like to see wikispeech deployed on Wikipedia, specifically in Arabic, and I would personally love to work for them in the future. Having Wikipedia with an Arabic Speech Synthesiser would encourage editors to correct and diacritise Arabic script on Wikipedia for it to be automatically read in a better way. This would make available more data for researchers to use in building their systems for Speech Synthesis and drive the technology forward.
Eric Luth
Project Administrator,
Wikimedia Sverige
+46765 55 50 95
eric.luth@wikimedia.se