Selective auditory attention in multi-participant environments

Abstract

System and method for managing audio in multi-participant environments enables simultaneous conversations through selective auditory attention. The system processes multimodal sensor data from wearable devices to determine attention patterns between participants. A probability matrix representing likely attention relationships is computed and used to dynamically adjust audio streams. The system maintains awareness of non-primary conversations while prioritizing active interactions. Machine learning techniques process participant interaction data to determine conversation groupings and predict attention changes. The system enables natural conversation flow in extended reality environments by selectively modifying audio gains based on detected attention patterns. Applications include professional meetings, educational settings, and social gatherings where multiple conversations occur simultaneously.

Dependencies are inferred from the claim text.