When
Mountain Standard Time (Arizona time)
Webinar video-recording
Slides
About the webinar
Arrival of ChatGPT has made interfaces like chatbots popular for asking questions about any topic and getting meaningful responses. With good prompts to guide them, chatbots can enable a purposeful conversation with relevant information from the broad knowledgebase and specific training materials. Such private Large Language Models (LLM), however, cannot readily interact with your private data sets and documents, thus limiting their utility for many researchers, educators, and students. Recently, many new open models like LLAMA and Mixtral that can run on your own computational infrastructure have become available. This allows users to build new analysis methods with these open LLM’s and integrate them with their private data sets and chatbots, enabling practical and time-saving uses.
In this webinar, presenters Nirav Merchant, Mithun Paul and Tyson Swetnam of UArizona show how you can use the no-cost NSF-funded JetStream2 A100 GPUs and tools from CyVerse to build your own private and secure GPT infrastructure. Learn how to have a conversation with your data!
About the presenters
Nirav Merchant is the Director of the Data Science Institute, Co-PI for NSF CyVerse and NSF Jetstream and leads the cyberinfrastructure team for the NSF & USDA funded National Artificial Intelligence Institute for Resilient Agriculture (AIIRA). Over the last two decades his research has been directed towards developing scalable computational platforms for supporting open science and open innovation, with emphasis on improving research productivity for geographically distributed interdisciplinary teams.
Mithun Paul is a Research Scientist and Educator at the Data Science Institute. Mithun has a PhD in Computer Science from the University of Arizona. He worked at the Information Sciences Institute of University of Southern California as a Research Scientist before joining the Data Science Institute. His research interests include artificial intelligence, natural language processing, quantum natural language processing and cyber security. He has widely published in top tier journals and conferences and holds several patents.
Tyson Swetnam is a Research Associate Professor of Geoinformatics and the Director of Open Science in the Institute for Computation and Data-enabled Insight at UArizona. He earned graduate degrees from the UArizona’s College of Agriculture, Life, and Environmental Science (CALES) in Watershed Management and Geographic Information Systems (GIS). His research covers a broad range of science and cyberinfrastructure applications where he collaborates with a diverse group of data science oriented projects in both Life and Earth Sciences and also leads multiple extramural research awards focused on data science, remote sensing of the environment, artificial intelligence and machine learning.