Intended Audience: Machine learning / predictive analytics practitioners who are interested in combining traditional approaches to building models with new generative AI technology.

Knowledge Level: Prior experience with data preparation and/or building supervised or unsupervised learning models, either using programming languages like R, Python, SAS, or GUI-based tools like KNIME, RapidMiner, WEKA, or others. Participants will have the opportunity to build generative-AI-based models using their favorite language or tool.

Workshop Description

With the emergence of AutoML and Generative AI as leading-edge and game-changing technologies, it’s natural for data scientists to ask the question, “How much can autoML do for me practically?”, and “Can generative AI actually build models for me?”. In this workshop, the question will be answered including both “pro” and the “con” sides of using autoML and Generative AI for predictive modeling. Examples with well-known datasets will be used to illustrate the concepts, and all prompts and code used in the workshop will be made available to attendees.

This workshop will include a hands-on portion.

6 hours: 4 1.5 hour parts

Part 1: Definitions

1. Different Flavors of AI

• a. Predictive Analytics -> Data Science -> Machine Learning -> AI

1. Generative AI: Deep Learning and LLMs
2. Why predictive AI is only a new marketing label for pre-existing technology
3. Automation: AutoML
4. Example Use Cases for illustration

Part 2: AutoML

1. What goes into AutoML?
2. What isn’t in AutoML?
3. How does AutoML connect to Predictive AI and Generative AI?
4. Live demo of an autoML framework and what it does for us.

Part 3: Build a predictive model using Generative AI (Live Demo)

1. Define the problem: Business Understanding
2. How Large Language Models can be applied for predictive use cases.

3a. Hands-on exercises, interactive, including prompts and solutions from ChatGPT 4o. Python code will be generated and run as well.

3b. Identify deficiencies of the analysis. Re-prompt

3c. Repeat

3d. Summarize results

Part 4: Additional Advanced Techniques for Model Building not Usually in AutoML Frameworks

1. Explanation of technique
2. GenAI Implementation of the technique, including Python code generation

Conclusions:

1. What does AutoML do well? What does it often miss?
2. What does Generative AI do well? What doesn’t it do well? Why or why not?

Schedule (schedule may change based on conference scheduling requirements)

8:30am Part 1
10:00 – 10:30am AM Coffee Break

10:30-noon: Part 2
noon – 1pm Lunch Break

1:00pm – 2:30pm Part 3
2:30pm – 3:00pm Afternoon coffee break

3:00pm – 4:30pm Part 4
4:30pm End of the Workshop

Instructor

Dean Abbott is Chief Data Scientist for Appriss Retail. He is an internationally recognized innovator in data science and machine learning with nearly four decades of experience solving problems in a wide range of applications. He is often included in lists of the most influential data scientists in the world, is author of Applied Predictive Analytics (2014) and a popular keynote speaker and instructor. He holds a BS in computational mathematics and a master’s in applied mathematics from UVA.