GPT for Social Research: How and Whether Large Language Models Can Help Social Scientists

Monday, April 3, 2:30pm - 4:00pm (EDT)

Link:

Time shown in -04:00 America, New York

The ability of GPT-3 and its "cousins" to generate texts for a broad range of tasks–from summarization to question-answering and information retrieval—via an intuitive natural language interface—may hold great promise for social science research. Already, researchers are using GPT-3 for tasks ranging from optical character recognition (OCR) correction to open-ended survey response measurement.

This session reviews what LLMs are, how they work, what they can do, and what the implications are of using these tools in academic social science research. There will also be a hands-on portion using OpenAI tools and supplementary libraries.

Participants who want to code along to the hands-on portion should be familiar with Python and must have an OpenAI account prior to the workshop.

The instructor, Dr. Musashi Jacobs-Harukawa, is a postdoctoral research associate at the Data-Driven Social Science Initiative. He received his doctorate at the University of Oxford, where he wrote his dissertation on applications of machine learning and natural language processing to studying political campaigns.

Please fill out the information below to register for the event.

Goingdone
Maybequestion_mark
Can't Goclear

Name

Attending in-person or Zoom?

Confirmed

A confirmation email is on its way. Check your spam folder if it doesn't show up.

Add to Calendar 2023/04/03 14:30:00 2023/04/03 16:00:00 America/New_York GPT for Social Research: How and Whether Large Language Models Can Help Social Scientists The ability of GPT-3 and its "cousins" to generate texts for a broad range of tasks–from summarization to question-answering and information retrieval—via an intuitive natural language interface—may hold great promise for social science research. Already, researchers are using GPT-3 for tasks ranging from optical character recognition (OCR) correction to open-ended survey response measurement.

This session reviews what LLMs are, how they work, what they can do, and what the implications are of using these tools in academic social science research. There will also be a hands-on portion using OpenAI tools and supplementary libraries.

Participants who want to code along to the hands-on portion should be familiar with Python and must have an OpenAI account prior to the workshop.

The instructor, Dr. Musashi Jacobs-Harukawa, is a postdoctoral research associate at the Data-Driven Social Science Initiative. He received his doctorate at the University of Oxford, where he wrote his dissertation on applications of machine learning and natural language processing to studying political campaigns. Robertson Bowl 016 Initiative for Data-Driven Social Science ddss@princeton.edu false 30

Response submitted

A confirmation email is on its way. Check your spam folder if it doesn't show up.

Response submitted

Thanks for letting us know.

Robertson Bowl 016

Initiative for Data-Driven Social Science, ddss@princeton.edu