Generative models \u2013 most prominently Large language Models (LLMs) for images and text \u2013 have been a major computing advance of 2023. For Ubiquitous Computing, these models offer tantalizing opportunities for personalization, data interpretation and more complex and meaningful interactions with humans.<\/p>\n
In this new era of AI there are many open problems, including how we assess them and how we deploy them both efficiently and safely. The Generative AI for Pervasive Computing Symposium<\/b> (GenAI4PC) aims to bring together both industrial and academic Ubiquitous Computing, researchers to explore these issues in a format that is focused on encouraging discussion and debate.<\/p>\n <\/div>\n For any information or queries, please contact the organizers at GenAI4PC@ubicomp.org<\/a><\/p>\n<\/div><\/div><\/div><\/div><\/div>Organizers<\/span><\/h3>\n
\n
Schedule<\/h3>\n<\/p>\n
\n\n
\n \nStart<\/th>\n End<\/th>\n Event<\/th>\n Talk Title<\/th>\n Author<\/th>\n<\/tr>\n<\/thead>\n \n 9:00<\/td>\n 10:30<\/td>\n NONE<\/td>\n <\/td>\n <\/td>\n<\/tr>\n \n 10:30<\/td>\n 11:00<\/td>\n Coffee \/ Networking<\/td>\n <\/td>\n <\/td>\n<\/tr>\n \n 11:00<\/td>\n 11:30<\/td>\n Introduction<\/td>\n <\/td>\n <\/td>\n<\/tr>\n \n 11:30<\/td>\n 12:00<\/td>\n Talk<\/td>\n Towards Enhanced Human Activity Recognition through Natural Language Generation and Pose Estimation<\/td>\n Nikhil, Kashyap, University of Washington<\/td>\n<\/tr>\n \n 12:00<\/td>\n 12:30<\/td>\n Talk<\/td>\n Taylorformer: Attention-based Generative Modelling for Random Processes including Time Series<\/td>\n Raghul, Parthipan, University of Cambridge<\/td>\n<\/tr>\n \n 12:30<\/td>\n 14:00<\/td>\n Lunch<\/td>\n [On your own]<\/td>\n <\/td>\n<\/tr>\n \n 14:00<\/td>\n 14:30<\/td>\n Panel<\/td>\n <\/td>\n <\/td>\n<\/tr>\n \n 14:30<\/td>\n 15:00<\/td>\n Panel<\/td>\n <\/td>\n <\/td>\n<\/tr>\n \n 15:00<\/td>\n 15:30<\/td>\n Talk<\/td>\n Speeding Up Question Answering Task of Language Models via Inverted Index<\/td>\n Xiang Ji, Boston university<\/td>\n<\/tr>\n \n 15:30<\/td>\n 16:00<\/td>\n BREAK<\/td>\n <\/td>\n <\/td>\n<\/tr>\n \n <\/td>\n <\/td>\n <\/td>\n <\/td>\n Rahul Islam<\/td>\n<\/tr>\n \n 16:00<\/td>\n 16:30<\/td>\n Talk<\/td>\n On the Benefit of Generative Foundation Models for Human Activity Recognition<\/td>\n Zikang, Leng, Georgia Institute of Technology<\/td>\n<\/tr>\n \n 16:30<\/td>\n 17:00<\/td>\n Talk<\/td>\n Revolutionizing Mental Health Support: An Innovative Affective Mobile Framework for Dynamic, Proactive, and Context-Adaptive Conversational Agents<\/td>\n Sang Won, Bae, Stevens Institute of Technology,<\/td>\n<\/tr>\n \n 17:00<\/td>\n 17:30<\/td>\n Talk<\/td>\n The first step is the hardest: Pitfalls of Representing and Tokenizing Temporal Data for Large Language Models<\/td>\n Dimitris, Spathis, Nokia Bell Labs<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n Contact<\/h3>\n