Many countries face the challenge of limited access to technology for the vast majority of people due to language barrier. One of those countries is Rwanda. The Digital Umuganda project proposes the first Kinyarwanda National Language Engine to provide access to technology for the Rwandan population.
Voice is natural, voice is human. With that in mind, we need to consider all sorts of things when it comes to accessing technology. Often, we face the challenge that not everyone speaks English, e.g. in Rwanda. Here, most people speak Kinyarwanda. On the contrary, most digital applications are built in English or French excluding a vast majority of the population. As a result, this creates a digital divide and excludes certain users from valuable services.
Giving Rwanda a voice
To reduce the digital divide and make technology accessible for all Rwandans, the Digital Umuganda project was born. DIGITAL UMUGANDA is an artificial intelligence and common digital infrastructure company currently focusing on voice technologies. It facilitates easy access to information and services in Kinyarwanda.
However, to create voice systems, developers need an extremely large amount of voice data. Most of the data used by large companies is not available to the majority of people. As a result, that stifles innovation. The use of the Common Voice platform in Kinyarwanda makes voice recognition open and accessible to everyone.
Originated from the Rwandan custom UMUGANDA
The name UMUGANDA is a practice that takes root from Rwandan traditional culture of self-help and cooperation. Members of the community come together to complete a difficult task such as building houses for the poor or road construction. Furthermore, the concept Umuganda originated to join forces and build a common digital infrastructure. In the spirit of Umuganda, the project organizes data collection events. Students for example can voluntarily donate and contribute on the Common Voice platform for the Kinyarwanda voice data set.
How does this look in a technical manner?
Digital Umuganda uses the Common Voice platform by Mozilla to collect Kinyarwanda voice data. It contains Kinyarwanda texts/phrases that can be read and recorded out loud by volunteer contributors. The goal is to record 1200 hours of Kinyarwanda voice data by the end of 2020.
Common Voice is part of Mozilla’s initiative to help teach machines how real people speak in the efforts to bridge the digital speech divide. Voice recognition technologies bring a human dimension to our devices. However, developers need an enormous amount of voice data to build them. Currently, most of that data is expensive and proprietary. Digital Umuganda wants to make voice data freely and publicly available and aims at collecting data representing the diversity of real people.
Projects by Digital Umuganda
This project addresses the challenge of accessing information and services as well as education and employment for people living with visual impairments. Digital Umuganda in partnership with Mozilla is building a Text to Speech (TTS) engine in Kinyarwanda. TTS will enable digital solutions to turn text into speech to ease accessing information. Additionally, TTS makes content available to people with learning disabilities, visual impairments or literacy challenges.
Challenges in coordinating responses, especially in information sharing and data gathering, put a burden on the health system. Digital Umuganda is working to respond to this challenges using artificial intelligence powered Chatbots. Those do not only ease information and knowledge dissemination in Kinyarwanda, but also put vital information in the hands of relevant authorities. As a result, this leads to a better coordinated two-way approach.
Digital Umuganda is still looking for people donating their voice to help building an open-source voice database. Read a sentence to help machines learn how real people speak. Donate your voice here.