C) Intro to Unsupervised Machine Learning (W38)

Note: Unsupervised Machine Learning Assignment Submission Deadline: Friday, 22 September 2023, 12:00

This topic includes 5 sessions as follows:

  • Introduction to Unsupervised Machine Learning (Mon, Sep 18th, 08:15-12:00): This session will dive into the foundational concepts and real-world uses of unsupervised machine learning (UML). As part of this, students will gain insights into various UML challenges. Furthermore, they will explore notable UML algorithms, including PCA, SVD, NMF, and an introduction to clustering via k-means.
  • UML 2: Recommendation & Similarity Search (Tue, Sep 19th, 12:30-16:15): This session will focus on the application of UML to recommendation systems and similarity search. Students will learn about the different types of recommendation systems, and they will explore how UML can be used to improve the accuracy and relevance of recommendations.
  • Building Recommender Systems with Gradio (Tue, Sep 19th, 16:30-18:15): This session will be a hands-on workshop where students will build their own recommender systems using the Gradio interface and UML methods. Students will work in teams to create a recommender system for a real-world problem.
  • Clustering extended: K-means & Hierarchical Approaches (Wed, Sep 20th, 08:15-10:00): This session will introduce the principles and applications of clustering. Students will learn about the different types of clustering problems, and they will explore some of the most popular clustering algorithms, such as K-means and hierarchical clustering.
  • **Demo session: Present your Streamlit App. All groups pitch their APP (5min - pitch). After that prices are given to the winning team and runner up. 🥳
  • UML Clustering Group Exercise (Wed, Sep 20th, 12:30-14:15): This session will be a hands-on exercise where students will cluster a real-world dataset using UML methods. Students will work in teams to identify the optimal number of clusters for the dataset, and they will analyze the characteristics of each cluster.