Welcome to CS 375

AI / Machine Learning
Spring 2025

Ken Arnold

As You Enter…

Discuss with your neighbors:

• What about AI are you hopeful about?
• What about AI are you concerned about?
• How might AI connect with Christian faith?

Each person should write one of each on the whiteboard.

My Story

• Cornell
• MIT Media Lab (+humanism)
• Luminoso
• Harvard
  • organizing things
  • writing and predictive text
• “Do justice with tech”?

Who are you?

• Introduce yourself to a few people around you. Try to meet someone you don’t know yet.
• Ideas to share (suggested by GitHub Copilot)
  • Your name
  • Your major
  • One thing you’re excited about this semester
  • One thing you’re concerned about this semester

Note: we have a few students participating as part of the Master’s of Data Science program.

What is this course?

This is a hands-on course on AI systems using machine learning, with a particular emphasis on deep neural networks.

Four Pillars of Modern AI

• Neural Computation: what are the “data structures and algorithms” of neural networks?
• ML Systems: what are common modules and APIs for ML systems?
• Learning Machines: How can systems learn from experience?
• Context and Implications: What can and should we use AI for? What broader questions should we ask?

Objectives, colloquially

You are effective at doing what’s just and right with AI.

So, you should be:

• self-directed, collaborative learners
• able to be immediately productive at some kinds of ML problems
• in the habit of practicing the integrity and perseverance needed to work in a team

Minimal Logistics

• Fill out survey and sharing signups
• Do the Perusall reading
• Buy the textbook

Also, reminder: CRA surveys.

Logistics

Meeting Log and Schedule

See Moodle for the link to the Google doc with a tab for each day.

Standards-Based Grading

• Why: to focus on learning, trust you, and give you flexibility.
• Details: let’s look at the syllabus together.

Notes:

• The Moodle gradebook is not yet set up to reflect this system. (I’m working on it.)
• I’m going to get this wrong the first time. Feedback is welcome!

Textbook

• Primary textbook: Deep Learning with Python, by François Chollet
  • if you buy the print version from Manning; you get the ebook for free
• Supplemental material from:
  • Understanding Deep Learning by Simon J.D. Prince (free online)
  • Deep Learning - Foundations and Concepts by Chris Bishop (free online, awkwardly)

Tech Stack

• Kaggle Notebooks (for running code on GPUs)
• Keras 3 (multi-backend, we’ll mainly use PyTorch)
• Hugging Face Transformers (for NLP and some CV)
• OpenAI API (for accessing commercial models)

AI Policy

• Use all types of AI maximally to help you learn.
  • Ask for analogies for complex concepts
  • Have it ask you questions to check your understanding
  • Talk about relationships between concepts
  • Use it to help write code
  • …
• Share your strategies with others (and we’ll discuss some together)

But:

• I don’t want to read any “slop”. (Show me your prompt instead!)
• Beware cognitive disengagement (the “feeling of learning”, thinking you understand because you can follow along)