Deep learning-based voice technologies offer a range of new opportunities to achieve the UN’s Sustainable Development Goals, like better health and reduced inequality. However, without the necessary capacity, technology, and training data, many countries and people who speak languages under-represented by current commercial solutions will miss out on the immense potential voice tech has to offer in the future. This is why in late 2019 Mozilla started collaborating with the Federal Ministry for Economic Cooperation and Development of the Federal Republic of Germany (BMZ) and its implementation organization Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ). Under the umbrella initiative "FAIR Forward - Artificial Intelligence for All" we want to open up voice technology for African languages.
Today we are very pleased to announce that Dr. Josh Meyer and Remy Muhire have joined as two new Mozilla Fellows dedicated to this project.
Dr. Meyer is currently the Lead Speech Scientist at Artie, Inc -- a Los Angeles-based Artificial Intelligence startup. He has conducted research at Mozilla as a National Science Foundation Fellow on the Machine Learning team. He holds a PhD from the University of Arizona, where he defended his thesis on Automatic Speech Recognition for Low-Resourced Languages in 2019.
Remy Muhire is an experienced community manager with an extensive background in software development. Remy is leading partnership development for early-stage voice data collection, advising the creation of open datasets in multiple languages and guiding the development of local products and use cases relevant to sustainable development goals.
The fellowship project is designed to support speech technology development for local languages in East Africa, beginning with Kinyarwanda (a widely-spoken Bantu language in Rwanda). This project will foster data collection efforts as well as provide technical advising for a team around the Rwandan AI startup “Digital Umuganda” to implement speech technology into an existing application. In particular, the project focuses on Automatic Speech Recognition. This is a useful technology which is commonly used in more developed countries, but it has not yet spread to East Africa. Speech recognition and speech synthesis are key to creating accessible technologies. People with vision impairments rely on speech synthesis to access the internet, read e-books, and navigate cities. From an economic standpoint, speech recognition creates substantial gains in efficiency, allowing humans to spend less time on tedious tasks.
Michelle Thorne, Senior Program Officer at Mozilla, says: “Josh and Remy make an ideal team to advance open voice technology grounded in local needs. Together, they bring expertise in state-of-the-art machine learning and under-resourced languages alongside technical community leadership in East Africa. Their fellowships will undoubtedly contribute to a healthier internet through open and sustainable approaches to voice technology.”
Mozilla’s fellowship program fuels the broader internet health movement. We provide funding, amplification, and networking opportunities to individuals devoted to privacy, inclusion, and openness online. Past Mozilla fellows have built browser extensions to protect users’ privacy; published localized Internet Health Reports; evangelized open-source tools for scientists; and more.
Some fellowships are exclusively operated by Mozilla, while others are operated alongside allies like Ford Foundation, Internet Society, Network Startup Resource Center, University of Dundee, and others. Mozilla’s fellowship work is made possible through work with our partners.