Abstract
The previously discrete technologies of Cyber-Physical/IoT Systems (CPS-IoT) and AI/ML have recently entered a tight, virtuous embrace. CPS-IoT technologies allow MEMS sensing and interventions in our physical, social, and urban spaces with unimaginable ubiquity. AI/ML methods allow sophisticated inferences and decisions to be made algorithmically using neural and neurosymbolic architectures, even from unstructured and high-dimensional spatiotemporal data, with uncanny performance. Together, they perform sophisticated perception-cognition-communication-action loops in diverse applications. After briefly discussing the current state of AI/ML methods for sensing in CPS-IoT systems, I will focus on the new opportunities and challenges for CPS-IoT emerging from Foundation Models, including Large Language/Multimodal Models, that have caused tremendous and broad excitement in AI. Based on preliminary research findings, the talk will provide insights into (i) the role that LLMs — with their world knowledge, ability to process sequences, and reasoning capabilities – can play in sophisticated MEMS sensor data analytics in CPS-IoT systems, (ii) the current capabilities and unresolved challenges of emerging foundation models focused on sensory data streams in the CPS-IoT domain.
Speaker(s): Prof. Mani Srivastava,
Agenda:
6:30 – 7:00 PM Registration & Networking
7:00 – 7:45 PM Invited Talk
7:45 – 8:00 PM Questions & Answers
Virtual: https://events.vtools.ieee.org/m/444309
MEMS Sensing in the Era of AI – Challenges & Opportunities from Foundation Models and LLMs
