Activity Detection, Mining, Scheduling, and Projection

This research is strongly tied to computational transportation, location awareness, and artificial intelligence based geo spatially aware pattern completion. This project includes various ways an individual’s activity might be determined whether through wearables, GPS location, images, etc. A secondary aspect of this is then applying large-scale data mining and machine learning of patterns of activities of individuals and using these to either help a user with time and geographically aware schedule optimization, or as a means of projecting temporal location and activity of a user for personalized recommendations. A major component of this work is how to do this in a privacy preserving way due to the potential large scale risk to privacy. Different aspects that have been part of this study have been improving machine learning for inferring from incomplete or missing data; and Generative Pretraining Transformer (GPT) applied to temporally geospatial aware activity scheduling projection for individuals and households.