Israeli startup listens to emotions, could change how we interact with machines

Listed by research firm Gartner as one of the five Cool Vendors in Human-Machine Interface for 2014, Beyond Verbal now claims close to 80 per cent accuracy rate in identifying emotions underlying human speech in controlled environments. “We have achieved this high rate of accuracy from 20 years of scientific and psychological research combined with findings from our database of close to a million emotionally tagged voice samples in more than 40 languages,” says Mor.

Emotions Analytics started when Beyond Verbal’s chief scientist Dr Yoram Levanon was trying to understand how babies manage to know exactly what their caretakers feel towards them. The physicist, with degrees in statistics and operation research, went on to study the physical modulations of air pressures that carry the sounds around us to look for the answer. Two decades of research with Neuropsychologist Dr Lan Lossos followed leading to 18 scientific discoveries and a battery of patents that now enables them to automatically extract human moods, attitudes and emotional characteristics from human voice intonations in real-time. Beyond Verbal was founded close to three years ago to commercialise the technology and has raised more than $7 million in funding so far.

Mor says understanding emotions applies to almost everything they do, from working out to understanding brands to providing services in a call center to playing games to interacting with our smartphone’s personal assistance. “Today Beyond Verbal’s technology is embedded in call centers, used by global brands via market research firms and powers third party mobile applications and enterprise solutions. We’ve merely just begun executing on our vision of enabling machines to sense, understand and interact with our emotions.”

Beyond Verbal currently offers its Emotions Analytics technology as an API targeting developers of emotionally enabled application, devices and solutions. Its demo apps Moodies and Empath record around 20 seconds of speech and analyse the emotions when it was spoken. Empath adds a layer of insight on how various activities over a period of time affect one’s emotional well being.

Beyond Verbal’s business model includes Cloud-based technology licensing, API geared towards wellness and fitness developers and leading operating systems for fast and painless implementation. “In opening our technology through an API over the cloud we look for partners that speak the same language and are on the same wavelength. We have started marketing Emotions Analytics usage packages to independent developers, established solution integrators and device manufacturers, says Mor, adding that in market research and service centers they have packages geared to these two separate verticals and priced by usage.

Mor’s goal is to endow machines with this emotions understanding and at the same time allow individuals to be more in tune with their own emotions. “On an operational level the company’s mission is to promote Emotions Analytics and lead a growing market of emotionally enabled solutions through the extension of Emotions Analytics into new market segments. We will bring emotional intelligence to B2B products, robots, consumer devices, and wearables.”