The funding will be used by the company to expand its global network of labellers as well as boost product development and hiring
Centaur Labs' founders (L to R): Tom Gellatly, VP of engineering, Erik Duhaime, CEO, and Zach Rausnitz, CTO (Credit: Business Wire)
Medical data labelling company Centaur Labs has secured $15m in Series A financing round to boost its efforts to label the world’s medical data and accelerate artificial intelligence (AI) development.
Led by Matrix Partners, the financing round witnessed participation from other funds such as Accel, Global Founders Capital, Susa Ventures, and Y Combinator.
It is also supported by individual investors such as WHOO founder and CTO John Capodilupo, One Medical founder Tom Lee and PillPack founder and CPO Elliot Cohen.
Centaur Labs will use the proceeds from the latest financing round to expand the firm’s global network of labellers, in addition to accelerating product development and hiring.
According to the company, healthcare companies need massive labeled datasets of medical images, videos, text or audio recordings for the training of AI algorithms.
Also, the efficacy of these algorithms is directly based on the accuracy of the underlying data labels.
To address this issue, Centaur has built a network of tens of thousands of medical students and professionals from more than 140 countries.
The network primarily labels data on Centaur’s gamified iOS app called DiagnosUs, which is designed to improve the skills of labellers.
Centaur has designed the app to judge labelers on their performance and reward the most accurate labellers.
The company will also gather multiple opinions on every case as well as intelligently combines those opinions into accurate labels.
Centaur Labs co-founder and CEO Erik Duhaime said: “AI learns like humans—by example—and to train an algorithm it takes thousands or even millions of examples. It is difficult to curate large medical datasets, and nearly impossible to source accurate labels from those with medical knowledge and specialized training.
“Our platform is built to support a wide range of specialized medical tasks, and to quickly scale to millions of labels.”