Nigeria 35 years old creates software that can understand and translate spoken African languages

Omolabake Adenle (35), is the founder of AJA.LA Studios, a startup building a platform of African language voice recognition and speech synthesis software. She holds a PhD in Bayesian Signal Processing from Cambridge University where she was a National Science Foundation Graduate Research Fellow and Tau Beta Pi Honors Fellow.

Abake’s innovation is software that can understand and digitize spoken African languages, and synthesize speech from African languages presented as digitized text.

Digitizing African languages in this way allows Africans to interact with hardware devices such as mobile phones, and digital services such as call-centre applications by speaking their local language.

The software can be integrated into a wide range of devices and third-party software applications.

 While voice recognition and speech synthesis software have been developed for various Western and Asian languages, there has been very limited commercial application or academic research for African languages.

The difficulty lies in modelling tonality present in most African languages, limited data resources for language modelling, and the challenge of modelling the plurality of African languages.

Using deep learning, Abake has developed complex algorithms and acoustic models for voice recognition and speech synthesis for two African languages.

She is currently working on 14 native and colonial African languages, with Swahili and Yoruba available for testing.

She is also currently working with various African retail banks and insurance companies on automating call-centre query resolution in local languages to help address issues of financial inclusion.

This innovation can open access to the benefits of technology for illiterate Africans who can access phones. It is also relevant to urban Africans who wish to interact with digital services in native or colonial African languages.