Jones Seminar: LLMs as Knowledge Orchestrators

Optional ZOOM LINK
Meeting ID: 946 1002 3166
Passcode: 428041

While the headlines around large language models (LLMs) are dominated by claims of omniscient AI, labor market revolutions, and copyright infringement, perhaps the most practical and longest lasting impact will be using LLMs as a platform for managing previously-siloed information within and across companies. This approach relies much more heavily on the language part of LLMs to understand user requests and to synthesize previously validated information. It provides an accessible user interface, enables and accelerates access to information, and avoids many of the accuracy and legal pitfalls of other LLM use cases. In this presentation, I will elaborate on this use case and discuss some of the architectures and strategies for using LLMs as a platform for knowledge management.

Hosted by Professor Geoffrey Parker.

About the Speaker(s)

Kurt DeMaagd
Chief AI Officer & Co-founder, Sight Machine

Kurt DeMaagd is Chief AI Officer and Co-founder at Sight Machine. He has developed ML applications including predictive maintenance, prescriptive energy optimization and waste reduction, decision support tools for quality, and schedule optimization. Kurt has presented on manufacturing AI/ML at conferences such as the Qatar Economic Forum, Automate, the American Society for Quality, multiple presentations at the ASQ Statistic Division conference, plus 30+ academic publications. Previously, Kurt was a professor at Michigan State University. He has a PhD in business administration from the University of Michigan, Ross School of Business, and a BS in computer science from Hope College.

Contact

For more information, contact Amos Johnson at amos.l.johnson@dartmouth.edu.