{"id":2983,"date":"2023-05-11T16:05:22","date_gmt":"2023-05-11T16:05:22","guid":{"rendered":"https:\/\/salarydistribution.com\/machine-learning\/2023\/05\/11\/startups-ai-slashes-paperwork-for-doctors-across-africa\/"},"modified":"2023-05-11T16:05:22","modified_gmt":"2023-05-11T16:05:22","slug":"startups-ai-slashes-paperwork-for-doctors-across-africa","status":"publish","type":"post","link":"https:\/\/salarydistribution.com\/machine-learning\/2023\/05\/11\/startups-ai-slashes-paperwork-for-doctors-across-africa\/","title":{"rendered":"Startup\u2019s AI Slashes Paperwork for Doctors Across Africa"},"content":{"rendered":"
\n

As a medical doctor in Nigeria, Tobi Olatunji knows the stress of practicing in Africa\u2019s busy hospitals. As a machine-learning scientist, he has a prescription for it.<\/p>\n

\u201cI worked at one of West Africa\u2019s largest hospitals, where I would routinely see more than 30 patients a day \u2014\u00a0 it\u2019s a very hard job,\u201d said Olatunji.<\/p>\n

The need to write detailed patient notes and fill out forms makes it even harder. Paper records slowed the pace of medical research, too.<\/p>\n

In his first years of practice, Olatunji imagined a program to plow through the mounds of paperwork, freeing doctors to help more patients.<\/p>\n

It\u2019s been a journey, but that software is available today from his company, Intron Health<\/a>, a member of the NVIDIA Inception<\/a> program, which nurtures cutting-edge startups.<\/p>\n

A Side Trip in Tech<\/b><\/h2>\n

With encouragement from med school mentors, Olatunji got a master\u2019s degree in medical informatics from the University of San Francisco and another in computer science at Georgia Tech. He started working as a machine-learning scientist in the U.S. by day and writing code on nights and weekends to help digitize Africa\u2019s hospitals.<\/p>\n

A pilot test during the pandemic hit a snag.<\/p>\n

The first few doctors to use the code took 45 minutes to finish their patient notes. Feeling awkward in front of a keyboard, some health workers said they prefer pen and paper.<\/p>\n

\u201cWe made a hard decision to invest in natural language processing and speech recognition,\u201d he said. It\u2019s technology he was already familiar with in his day job.<\/p>\n

Building AI Models<\/b><\/h2>\n

\u201cThe combination of medical terminology and thick African accents produced horrible results with most existing speech-to-text software, so we knew there would be no shortcut to training our own models,\u201d he said.<\/p>\n

\"Tobi<\/a>
Tobi Olatunji<\/figcaption><\/figure>\n

The Intron team evaluated several commercial and open-source speech recognition frameworks and large language models<\/a> before choosing to build with NVIDIA NeMo<\/a>, a software framework for text-based generative AI<\/a>. In addition, the resulting models were trained on NVIDIA GPUs in the cloud.<\/p>\n

\u201cWe initially tried to train with CPUs as the cheapest option, but it took forever, so we started with a single GPU and eventually grew to using several of them in the cloud,\u201d he said.<\/p>\n

The resulting Transcribe app captures doctors\u2019 dictated messages with more than 92% accuracy across more than 200 African accents. It slashes the time they spend on paperwork by 6x on average, according to an ongoing study Intron is conducting across hospitals in four African countries.<\/p>\n

\u201cEven the doctor with the fastest typing skills in the study got a 40% speedup,\u201d he said of the software now in use at several hospitals across Africa.<\/p>\n

Listening to Africa\u2019s Voices<\/b><\/h2>\n

Olatunji knew his models needed high quality audio data. So, the company created an app to capture sound bites of medical terms spoken in different accents.<\/p>\n

To date, the app\u2019s gathered more than a million clips from more than 7,000 people across 24 countries, including 13 African nations. It\u2019s one of the largest datasets of its type, parts of which have been released as open source to support African speech research.<\/p>\n

Today, Intron refreshes its models every other month as more data comes in.<\/p>\n

Nurturing Diversity in Medtech<\/b><\/h2>\n

Very little research exists on speech recognition for African accents in a clinical setting. So, working with Africa\u2019s tech communities like DSN<\/a>, Masakhane<\/a> and Zindi<\/a>, Intron launched AfriSpeech-200<\/a>, a developer challenge to kickstart research using its data.<\/p>\n

Similarly, for all its sophistication, medtech lags in diversity and inclusion, so Olatunji recently launched an effort that addresses that issue, too.<\/p>\n

Bio-RAMP Lab<\/a> is a global community of minority researchers working on problems they care about at the intersection of AI and healthcare. The group already has a half dozen papers under review at major conferences.<\/p>\n

\"Olatunji<\/a>
Olatunji presents his ideas at NVIDIA\u2019s Santa Clara campus in a meeting kicking off an initiative to make AI accessible for all.<\/figcaption><\/figure>\n

\u201cFor seven years, I was the only Black person on every team I worked on,\u201d he said. \u201cThere were no Black scientists or managers, even in my job interviews.\u201d<\/p>\n

Meanwhile, Intron is even helping hospitals in Africa find creative ways to acquire the hardware they need. It\u2019s another challenge on the way to opening up huge opportunities.<\/p>\n

\u201cOnce healthcare data gets digitized, you unlock a whole new world for research into areas like predictive models that can be early warning systems for epidemics \u2014 we can\u2019t do it without data,\u201d Olatunji said.<\/p>\n

Watch a masterclass<\/a> (starting at 20:30) with Olatunji, HuggingFace and NVIDIA on AI for speech recognition.<\/p>\n<\/div>\n","protected":false},"excerpt":{"rendered":"

https:\/\/blogs.nvidia.com\/blog\/2023\/05\/11\/ai-africa-doctors-paperwork\/<\/p>\n","protected":false},"author":0,"featured_media":2984,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[3],"tags":[],"_links":{"self":[{"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/posts\/2983"}],"collection":[{"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/types\/post"}],"replies":[{"embeddable":true,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/comments?post=2983"}],"version-history":[{"count":0,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/posts\/2983\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/media\/2984"}],"wp:attachment":[{"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/media?parent=2983"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/categories?post=2983"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/tags?post=2983"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}