Workshop on Generative AI: An Innovative Research Tool

Friday, September 22 & Friday, September 29
Vance 112

The Mack Institute is pleased to announce a two-day workshop focusing on Generative Pretrained Transformers (GPT) as a major innovation in academic research.

Day 1 will be devoted to coding-free research activities while Day 2 will cover coding in Python with Github’s Copilot. The ideal participant is the one who wants to employ Large Language Models (LLMs) in their research projects, has some prior experience with coding but lacks proficiency with Python.

The workshop is open to Wharton faculty, PhD students, predocs and postdocs. Preference will be granted to active researchers, with particular attention given to those who are currently receiving research funding from the Mack Institute.

In order to ensure an optimal learning experience, the class size will be limited to 40 participants. Attendance on both days is highly desired but, if the space allows, we will accommodate those who commit to Day 1 only.

Click Here To Register


Agenda

Friday, September 22 – Day 1, No Coding Needed

Unless specified, units are taught by Lennart Meincke & Edward Zhang

8:00 – 9:00: Breakfast & Networking

9:00 – 9:30: Overview, Introductions, and Discussion on the Implications of AI for MBAs – Christian Terwiesch

9:30 – 10:15: GPT for Innovation and Ideation, Mack White Paper on Innovation, Research Approach – Christian Terwiesch & Lennart Meincke

10:15-10:30: Introduction of Participants

10:45-12:00: LLMs 101: Models, Words and Tokens

12:00-1:00: Lunch

1:00-2:30: ChatGPT Beyond the Prompt Box: Playground and Plug-ins

2:30 – 2:45: Coffee Break

2:45 – 3:45: Testing Prompts, Model Parameters, API Key

3:45 – 4:15: Other Forms of Generative AI: Midjourney and Dall-E

End of Day 1: Stay late if you need help with GitHub, Python, Visual Studio Code or API setup

Note: For students making good progress, self-taught material for web scraping using Beautiful Soup will be provided.


Friday, September 29 – Day 2, Start Improving Your Code With GPT

Unless specified, units are taught by Lennart Meincke & Edward Zhang

8:00 – 9:00: Breakfast & Networking

9:00 – 10:30: Keynote: Few-shot Learning, Prompt Engineering, State of the Art Research, Model Limitations, LLM Governance – Chris Callison-Burch

10:30 – 11:00: Homework Presentations

11:00 – 12:00: Coding in the Age of LLMs: GitHub Copilot

12:00 – 1:00: Lunch Break – Support for attendees with technical issues

1:00 – 1:30: From Playground to VS Code: Automating GPT

1:30 – 2:30: Tutorial on Automating CSV Analysis Using OpenAI API

2:30 – 2:45: Coffee Break

2:45 – 3:45: Tutorial on Automating CSV Analysis Using OpenAI API (continued)

3:45 – 4:15: Audience Q&A, Discussion on Integrating AI into Research Needs, Wrap Up

End of Day 2: Open Discussion for Questions on Using GPT for Research. Come Chat with Us!

Click Here To Register

Instructors: Christian Terwiesch (Andrew M. Heller Professor at the Wharton School; Professor of Operations, Information and Decisions; Professor of Health Policy, Perelman School of Medicine; Co-Director, Mack Institute of Innovation Management; Department Chair), Chris Callison-Burch (Associate Professor, Computer and Information Science), Lennart Meincke (Instructor, University of Pennsylvania), Edward Zhang (Instructor, University of Pennsylvania)