Lecture: Machine Learning for Graphs and Sequential Data

This course (IN2323) builds upon the knowledge you gained in the lecture Machine Learning (IN2064). It provides advanced learning principles and covers more complex data domains.

Information

  • Lecture/Exercise: Wednesday and Thursday 14:00-16:00
  • We will primarily communicate through Piazza, while using Moodle to distribute the learning material. The password will be made available on Moodle at a later date.
  • Required knowledge: Content of our Machine Learning lecture

All announcements will be made on the Piazza forum, which can be accessed via the link on the course's moodle page.
Please do not send any questions about organizational matters via e-mail.
If you have problems accessing the Moodle course, contact d.luedke [at] tum.de.

Additional to MLGS, we have introduced a new course on Deep Generative Models this year. If you are interested, please have a look on the course website: DGM 

 

Tentative list of topics MLGS

  1. Sequential data
    • AR and Markov Chains
    • Hidden Markov Model
    • Sequence models (Recurrent Neural Networks, Transformers, etc.)
    • Embeddings (Word2Vec)
    • Temporal Point Processes
  2. Graphs
    • Laws and Patterns
    • (Deep) Generative Models for Graphs
    • Spectral Methods
    • Clustering
    • Embedding and Ranking (e.g., PageRank)
    • Community Detection
    • Node/Graph Classification
    • Label Propagation
    • Graph Neural Networks (expressivity, oversmoothing, invariances etc.)
  3. Robustness
    • Adversarial attacks
    • Adversarial training
    • Exact robustness verification
    • Relaxed robustness certification (Convex relaxation, Lipschitzness, Randomized smoothing)
    • GNN robustness