NISHA: NTT - Idiap Social beHavior Analysis Initiative

Introduction

Idiap's Social Computing Group and NTT's Communication Science Labs are jointly investigating methods to automatically infer social constructs in everyday interactions, using nonverbal behavioral cues as observations.

We are investigating two settings: face-to-face teamwork in organizations and online social video. In both of them, the nonverbal cues that people generate (nods and smiles, postures and tone of voice) can be used to understand key aspects of their state of mind and personal traits.

Combining ubiquitous sensing devices and machine learning methods, NISHA aims to detect behavioral patterns, analyze their properties, and use this knowledge to build systems that enable self and group awareness.

Keywords: automatic behavioral analysis, social computing, face-to-face communication, social video.