Computers can beat humans at games like chess and Go. They are obviously faster than us at calculations and numerical analysis. But they can’t read people’s emotions as well as we can, or do much of anything useful with that information. Not yet, anyway.
But researchers and companies have been trying to close that gap. Now, Affectiva, a Waltham, MA-based tech company, has raised $14 million in new funding to advance its emotion-sensing technology and products. The news was reported by TechCrunch.
The round was led by Fenox Venture Capital, and it brings Affectiva’s total venture-capital haul to about $34 million. Prior to this, the company’s most recent financing was a $12 million Series C round in 2012, from Horizons Ventures, Kleiner Perkins Caufield & Byers, WPP, and Myrian Capital.
Affectiva was spun out of the MIT Media Lab in 2009 by scientists Rosalind Picard and Rana el Kaliouby. Picard is no longer with the company, and el Kaliouby is now its CEO. The company started out measuring emotions using two different systems: webcams (computer vision) and wearable sensors. In recent years, Affectiva has focused on the vision-based approach, using machine learning techniques to analyze people’s facial expressions and non-verbal cues; the main applications have been in advertising, marketing, and gaming. (On the wearables side, Picard co-founded a separate healthcare-focused company, Empatica, where she is chief scientist.)
Within artificial intelligence, the field of emotion recognition has seen increased interest among startups, investors, and big companies. In January, Apple acquired San Diego-based Emotient, a facial-analysis tech company. Another startup developing emotion-recognition software is Eyeris, based in the Bay Area. It’s not hard to imagine that companies like Facebook, Google, and Amazon would be interested in understanding consumers’ reactions to online content (and their interactions with each other). The question is what the companies will do with that information once it becomes available and reliable.
