Plan of the workshop

[kvarada]

Activity

Start with an activity where we ask them to frame their problem which they think might be suitable for ML.

• What are you trying to accomplish? State the goal of the solution you are interested in developing
• What’s the current approach? How does a human solve the problem?
• What kind of data is available to you, or could be available? What are some example features?

One of the main learning objectives of the workshop will be to determine whether your goal is best addressed using supervised machine learning, inferential statistics, unsupervised learning, deep learning, generative AI, or a non-ML solution.

Introduction to supervised ML (~60 mins)

• inference vs. prediction
• types of ML
• types of data
• regression vs. classification
• decision trees
• ML fundamentals (generalization, data splitting, overfitting/underfitting, fundamental tradeoff, golden rule )
• resources: 571 course info, 571 ML fundamentals

Break (~5 mins)

Supervised ML models and pipeline (~30 mins)

• Different classes of supervised ML models: tree-based models, linear models, analogy-based models (time permitting)

Intro to deep learning/image classification (~30 mins)

• High-level intro to neural networks (e.g., stacked logistic regression modules + non-linearity)
• pre-trained models, transfer learning

Break (~5 mins)

Intro to unsupervised learning (~20 mins)

• Two main types of unsupervised learning: clustering and dimensionality reduction
• Use cases of clustering
• Latent features using dimensionality reduction

Intro to LLMs (~20 mins)

• What are LLMs?
• Examples of LLM use cases
• Prompt engineering