Using AI to Translate Baby Cries? A New App Is on the Way

Had Apolline Deroche, MFin ’22, decided to attend medical school instead of studying engineering at the École des ponts ParisTech, she probably would have become an obstetrician.

“Early childhood health and development is so fascinating,” says Deroche, founder and CEO of Cappella, a startup that uses artificial intelligence and machine learning to generate soothing sounds for infants and translate their cries for parents.

Cappella, which Deroche developed during her time in the MIT Sloan Master of Finance program, is the culmination of her interests in engineering, music and sound, and child health and psychology. “I didn’t do that on purpose,” she explains, “but I think this is why I’m so passionate about Cappella even after working on it for over a year now.”

The cry detection and translation app is still in early testing, but Deroche, who just completed the Techstars accelerator, is excited by the current possibilities and potential future iterations of AI and audio technologies in the health care space.

“I want to make it my life’s work to use technology to improve people’s lives,” she says.

Turning ideas into actions

In her first year, Deroche set out to do this by enrolling in the Entrepreneurship Lab (E-Lab), one of over 15 Action Learning labs in which students embody the Institute motto of “mens et manus,” or “mind in hand,” by bringing their classroom learnings to address real management opportunities and challenges presented by host companies.

“Students have the opportunity to work with cutting-edge startup ventures on their most pressing issues—gaining invaluable exposure to entrepreneurship and direct access to founders solving some of the world’s most pressing problems,” says Kit Hickey, MBA ’18, (Senior Lecturer and Entrepreneur in Residence).

Becca Souza (Director, Action Learning) adds, “Our goal for students is to reflect on these experiences and use what they learn for their personal and professional growth.”

Deroche’s E-Lab group partnered with BioSens8, a biotechnology company specializing in small, noninvasive medical devices. Every week, the student team met with BioSens8 CEO and Co-Founder Uroš Kuzmanović to identify unmet needs and value propositions for a progesterone biosensor the startup was working on.

“Those weekly discussions were truly a highlight for me. The data and ultimate report we put together laid the groundwork and confidence for BioSens8's entry into women's health,” says Kuzmanović. “I made connections with brilliant people on the team, like Apolline, whom I keep in touch with today. Seeing her take the next step to create impact herself with Cappella has been exciting to follow.”

A longtime admirer of entrepreneurship and those who practice it, Deroche quickly realized that she too could do what Kuzmanović and other startup founders like him did.

“If he can do it, I can do it, too,” she remembers thinking at the time. “I had no idea how to start from a blank page at the time, so I feel like Entrepreneurship Lab really taught me how to start from a blank page.”

Additional entrepreneurship classes like Entrepreneurial Founding and Teams with Hickey and Erin Scott (Senior Lecturer, Technological Innovation, Entrepreneurship, and Strategic Management), and New Enterprises reinforced Deroche’s newfound determination and helped her cultivate the ideas that would eventually become Cappella.

“I’m so grateful to MIT Sloan, because I think that, otherwise, I don’t know if I would have ever given my dreams a chance,” says Deroche, who returned to E-Lab last spring with Cappella as a host company.

The new industrial revolution

From predicting protein structures to accelerating drug discovery and testing, recent advances in machine learning and other AI technologies are rapidly fostering what Vijay Pande, PhD ’95, general partner at Andreessen Horowitz, calls a “new industrial revolution” in health care.

The potential for the AI-enabled discovery and development of new diagnostics, treatments, and cures in this new industrial revolution is what excites Deroche most about her early work at Cappella, as well as what the future of AI and audio technology might hold.

Consider the recent example provided by the MIT Auto-ID Laboratory and recent health care startups using AI and a database of cellphone-recorded coughs to detect and diagnose COVID-19 and other serious respiratory ailments.

“Instead of getting the nose swab, you just have to record yourself coughing. This is a tiny example, but there is so much more in store for health care, and that’s the most exciting thing,” says Deroche.

For Cappella, this revolution has meant research partnerships with hospitals and friendly conversations with the big brands in the baby monitor business, which could see a projected compound annual growth rate of 9.45% from $1.48 billion in 2023 to $2.32 billion in 2028. Many of these discussions are early and ongoing, but Deroche’s enthusiasm is seemingly limitless.

“We want to advance the state of this technology—baby cry translation and recognition—because we know there’s so much more we can do with it,” she says. “We’re building on what has already been done and providing researchers with the opportunity to do more research and develop more advancements in medicine and well-being—all through a baby’s cries.”

As focused as Deroche has been on research applications and B2B, however, she has never lost sight of an incredibly powerful (and interested) group of consumers: parents. A demo is currently available, as is a waitlist for the cry detection and sound generator app, but parents are regularly expressing their excitement while also asking when the app will be available.

“The response has been really great,” says Deroche, “and we’re super excited to finally come out with the actual product.”