Online Seminar Computer Science - Dr. Stephanie Tan, Technical University Delft

Title: Towards socially intelligent computing: sensing and modeling human social behaviors

Abstract:

In recent years, new domains such as social signal processing have emerged at the intersection of computer science, human behavioral modeling, and robotics. The aim of these fields is to achieve machine perception of social intelligence, such as understanding behavioral cues of humans (e.g., body language) and complex social relations and attitudes (e.g., dominance, rapport). Challenges towards building such systems include data acquisition with appropriate sensing capabilities for capturing in-the-wild human data, as well as modelling approaches that account for data from multiple modalities (vision, audio, motion). In this talk, I will first introduce my approach to some of the open questions in obtaining ecologically valid and high-fidelity human behavior data. Then, I will discuss data-driven approaches for automatic inference of human behavioral cues and scene understanding in face-to-face interactions. In particular, I will highlight the machine learning frameworks that I used to address issues related to context modeling and multimodal data sparsity and quality. I will close by discussing outlooks and application of socially intelligent computing to real- life settings, particularly in (i) improving human-human interactions with automatic recognition of micro-behaviors in workplace settings and (ii) improving human-robot interactions by maximizing physical safety, adoption, and trust in industrial settings (e.g., warehouses).