Effective Prompt Engineering for AI-Powered Research Chat Tools
Alternative Title
Effective Prompt Engineering for Artificial Intelligence (AI)-Powered Research Chat Tools
Contributor
University of Central Florida. Faculty Center for Teaching and Learning; University of Central Florida. Division of Digital Learning; Teaching and Learning with AI Conference (2024 : Orlando, Fla.)
Location
Mangrove
Start Date
23-7-2024 2:45 PM
End Date
23-7-2024 3:45 PM
Publisher
University of Central Florida Libraries
Keywords:
Prompt engineering; AI research tools; Information retrieval; Large language models; Source citation
Subjects
Prompting (Education); Information retrieval--Research; Natural language generation (Computer science); Artificial intelligence--Research; Question-answering systems
Description
Retrieval Augmented Generation enables a rapidly expanding class of AI-powered research tools (e.g. Perplexity, Elicit, SciSpace) to conduct searches of articles and/or the broad Internet and then use the results of those searches to cite authentic sources and develop detailed answers. In this session we demystify this process by explaining how these tools translate user-supplied prompts into searches, how the retrieved information is processed by the underlying Large Language Model, and how the LLM cites sources. We then explore how to use this knowledge to improve the quality of the retrieved sources and the relevance of the answer.
Language
eng
Type
Presentation
Format
application/vnd.openxmlformats-officedocument.presentationml.presentation
Rights Statement
All Rights Reserved
Audience
Faculty, Students, Librarians
Recommended Citation
Coleman, Jason and Olsen, Livia, "Effective Prompt Engineering for AI-Powered Research Chat Tools" (2024). Teaching and Learning with AI Conference Presentations. 101.
https://stars.library.ucf.edu/teachwithai/2024/tuesday/101
Effective Prompt Engineering for AI-Powered Research Chat Tools
Mangrove
Retrieval Augmented Generation enables a rapidly expanding class of AI-powered research tools (e.g. Perplexity, Elicit, SciSpace) to conduct searches of articles and/or the broad Internet and then use the results of those searches to cite authentic sources and develop detailed answers. In this session we demystify this process by explaining how these tools translate user-supplied prompts into searches, how the retrieved information is processed by the underlying Large Language Model, and how the LLM cites sources. We then explore how to use this knowledge to improve the quality of the retrieved sources and the relevance of the answer.
