Researchers at MIT Media Lab are working on an automated personal-computer-based system designed to help people improve interpersonal and conversational skills.
The software is called MACH, short for My Automated Conversation coacH. It makes use of a computer-generated onscreen face to simulate interactive conversations. It performs facial, speech, and behavior analysis and synthesis to emulate human-to-human responses. At the end of the session it provides the user with feedback on their performance.
It has already been shown to improve job interview techniques and other potential uses are coaching in public speaking and dating. The research team in the Affective Computing Group of MIT Media Lab is expanding the technology with the intention of treating people with Asperger syndrome and PTSD (post traumatic-stress disorder) and helping overcome social phobias.
In this video MIT Media Lab doctoral student M. Ehsan Hoque, who led the research, explains that the system was designed to meet a widespread requirement of people with social phobias of having:
"some kind of automated system so that they can practice social interactions in their own environment. … They desire to control the pace of the interaction, practice as many times as they wish, and own their data.”
Using a webcam the program can analyze facial expressions, including smiling and head gestures while its voice recognition system analyzes not only what you say but also how you say it and notes non-verbal vocalizations. As feedback you watch the video side-by-side with the analysis of your behaviour.
The effectiveness of MACH was assessed through a week-long trial with 90 MIT undergraduates. Students who interacted with
MACH were rated by human experts to have improved in overall interview performance, while the ratings of students in control groups did not improve. Post-experiment interviews indicate that participants found the interview experience informative about their behaviors and expressed interest in using MACH in the future.
More details are given in a paper to be presented at 2013 International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp 2013).