namu

Research

We study how machines hear, understand and answer in the languages of the Sahel

ProgrammesFive long-running research efforts, each with its own questions and milestones

View programmes
Full-duplex conversation in Hausa

Hira

Speech models that listen while they speak, built for Hausa and judged against how Hausa speakers actually take turns.

Can a machine listen while it speaks, in Hausa, and take turns the way Hausa speakers do?

Explore Hira
A consented Hausa speech commons, Niger first

Rumbu

Hausa speech collected with consent and fair pay, from the people and places public datasets leave out.

What would Hausa speech data look like if it were collected with the people it represents?

Explore Rumbu
Zarma from the ground up

Zarma

The first open, consented speech resources and honest baselines for Zarma, spoken by about six million people along the Niger.

What does it take to build speech technology for a language that almost none exists for?

Explore Zarma
From raw audio to trustworthy data

Tankaɗe

The methods that turn radio archives, field recordings and crowd audio into Hausa and Zarma data a model can learn from and a reader can trust.

Which choices in cleaning speech data change what a model learns?

Explore Tankaɗe
Honest measurement for Hausa and Zarma speech

Maʼauni

Test sets, text normalisation, listening protocols and benchmarks that make a Hausa or Zarma result mean something.