As alluded to in the sidebar, this blog is about what it will take to achieve social learning machines. Some intuitive examples of social learning machines come from Hollywood (The Jetson’s Rosie, and Star War’s R2D2 and C3PO). These machines are general purpose, they co-exist with people in a human culture and successfully interact in human environments. And importantly, they have the ability to learn about their environment and new challenges that arrise by interacting with the people around them.
While this is currently science fiction, many industrial and academic research labs are working towards this vision of robots and machines as helpful assistants to humans. Social learning is an important challenge in this effort and requires a cross disciplenary effort in the fields of Machine Learning and Human-Computer Interaction. My academic research is Socially Guided Machine Learning (SG-ML), and is at the intersection of these fields. My work aims to analyze, critique, and augment current Machine Learning algorithms and techinques in order to make them more amenable to learning from an everyday human partner. The purpose of this blog is to bring together thoughts from my own research and related works in industry and academia. And in the end…we’ll all get our robots!
About the Author:
My name is Andrea Thomaz, and I am an Assistant Professor in the Interactive Computing department at Georgia Tech, where I direct the Socially Intelligent Machines research lab. Previously, I completed my Ph.D. with Cynthia Breazeal in her Robotic Life Group at MIT. Over the past 7 years I have been researching the ways in which Aritificial Intelligence and Machine Learning and Human-Computer (Robot) Interaction can be mutually beneficial (i.e., HCI can leverage Machine Learning and Machine Learning can leverage human interaction). For more information about me and my research visit my homeage at GaTech, and the previous one at MIT.