Instructors: Edward Kim, Cohere AI; Jason Hattrick-Simpers, University of Toronto

In this tutorial, attendees will get a hands-on introduction to leveraging large language models (LLMs) for accelerating materials research. 

The Cohere platform will be used as an exemplar access point to leveraging the capabilities of large language models, including: 

• Grounding responses against scientific facts or documents
• Querying a LLM with followup questions about a scientific document
• Probing the reasoning and level of consistency in LLMs
• Using LLMs to “glue” together pieces of research software
• Extracting and transforming data from scientific sources

This tutorial will include direct natural language interfaces with models, as well as API-driven programmatic interfaces with models. 

Attendees will learn from this tutorial how they can best leverage LLMs in accelerating their own scientific work, and how to assess when LLMs are the right tool for the job.

Tutorial Schedule