Back to Blog

Social Smarts for Robots

By Rohit Sharma, January 10, 2018

Share on TwitterShare on FacebookShare on LinkedIn

Automation and AI discussions tend to bring up widely varying viewpoints. While some countries are optimistic about how automation and AI can aid humans, perspectives in the Valley and U.S. in general seem to swing to fear. This fear can at times be driven by economics, but it’s also related to social implications.

How does one work harmoniously with a robot? Will robots ever be able to pick up on social cues?

In the case of industrial automation where robots are programmed to perform one task repetitively, there isn’t a need for social awareness or the ability to recognize human collaboration. Programmed behaviors are predictable… controllable. But what about when robots must coexist in the same workspace with humans and perform collaborative actions alongside human workers? Social learning abilities are required in order for robots to succeed in those scenarios.

Today, we are proud to announce our investment in Diligent Robotics, a human-centered robotics company co-founded by Andrea Thomaz and Vivian Chu. Diligent is looking to teach robots how to work in a collaborative manner with humans — an exciting new domain and a difficult problem to solve.

The Diligent team begins their journey by producing an autonomous hospital service robot meant to assist clinical staff with menial tasks so they can spend more time on patient care. From there, Diligent will explore other applications for socially intelligent robotics.  

Hospitals are good grounds for testing this type of social technology because they provide a structured environment in terms of layout (uniform hallways, similar room settings, etc.) that robots can learn while figuring out how to deal with the infinite variations of human behavior.

In the case of Diligent’s hospital service robots, it will be easy for them to learn how far they must travel to fetch a new IV bag, but it’s a much more difficult problem to teach them how to travel down the hallway in a way that does not intrude upon the pathways of doctors, nurses and patients throughout the hospital. Decisions like when to step aside to let someone pass through a doorway are simple for humans, but complex for social robots.

So how do you teach robots how to aid humans in a socially adept, scalable way? If there is any team primed to find the answer to this question, it is team Diligent. Andrea and Vivian are both deeply experienced roboticists with impressive academic backgrounds and a love for tinkering. 

In the video above, Andrea explains the intricacies of robots that can collaborate with humans and understand social cues. 

Andrea has 15 years of experience researching social robots and machine learning. She currently leads the robotics lab in the electrical and computer engineering department of University of Texas at Austin and completed her Ph.D. thesis on socially guided machine learning at MIT. She has been recognized as a leader in her field by the National Academy of Science, Popular Science Magazine and the President’s Council of Advisors on Science and Technology.

Vivian has over six years of experience programming mobile manipulation platforms and applying machine learning algorithms within various robotic platforms. Prior to Diligent Robotics, she was a user experience researcher at Google X where she conducted user studies and prototyped delivery mechanisms for Project Wing, Google’s delivery drones experiment. In 2016, recognized Vivian as one of the leading women in the robotics industry.

Diligent Robotics

Diligent Robotics Co-Founders Vivian Chu and Andrea Thomaz pose with an early prototype of the robot they are building to work alongside hospital staff. They will kick off a pilot program in U.S. hospitals this year. Learn more in the Wall Street Journal.

Welcome to the True community, Andrea and Vivian! We’re thrilled by your curiosity and drive to tackle this interesting problem.