On January 16, 2024, we hosted our first AI Interest Group Workshop! It was a wonderful chance to learn about some of the projects that library staff are working on involving Artificial Intelligence.
Automating Map Georeferencing using Machine Learning
Michael Shensky, the Head of Research Data Services at UT Austin, shared an overview of an ongoing machine learning project that automates georeferencing for scanned maps within the UTL collection. With a growing collection of over 60,000 maps, automating the process of georeferencing and metadata creation will be crucial for efficiently managing the workflow in this area. The slides from Michael’s presentation can be found here.
Experimenting with Large Language Models
Hannah Moutran, a Graduate Research Assistant studying Implementations of AI for UT Libraries, led staff through some practical applications of Large Language Models (LLMs). The presentation began with an overview of commercial platforms and OpenAI’s GPTs. It covered open source models, Hugging Face, and LLM evaluation metrics, and included a brief demonstration of how to easily run open source models both locally and using a cloud computing service. The presentation ended with a quick overview of how to create a simple Python chatbot. Slides from the presentation are available here.
Thank you to our presenters and the attendees – we look forward to continuing to learn and grow through our future sessions! Our next meeting, planned for late February, will focus on the Ethics of AI for Libraries, Archives, and Museums. If you are interested in attending, please fill out a form on our Events page.