This website is the written documentation of the Seminar Principles of Data Mining and Learning Algorithms – „Learning Theory“ MA-INF 4209, taught at the University of Bonn during the Winter Termin 2018/2019 by Pascal Welke & Michael Kamp.

The seminar is based on the book „Understanding Machine Learning“ by Shai Shalev-Shwartz and Shai Ben-David. Below, you find the write-ups of our sessions.

Session | Presenter | Book Chapter | Link | Screencast |
---|---|---|---|---|

1 | Michael Kamp | 1 & 2 | Introduction and Gentle Start | Screencast |

2 | Mikhail Borisov | 3 & 4 | APAC Learning and Uniform Convergence | Screencast |

3 | Lukas Drexler | 5.1 | No Free Lunch Theorem | Screencast |

3 | Lukas Drexler | 5.2 | Error Decomposition | Screencast |

3 | Lukas Drexler | 6 | VC-Dimension | Screencast |

4 | Oliver Kiss | 6 | Addition VC-Dimension | Screencast |

4 | Oliver Kiss | 7 | Nonuniform Learnability | Screencast |

4 | Oliver Kiss | 6 & 7 | Q&A | |

5 | Maximilian Thiessen | 7 | MDL and Other Notions of Learnability | Screencast |

5 | Pascal Welke | 7 | An Alternative Proof for Thm 7.2 | |

5 | Maximilian Thiessen | 8 | The Runtime of Learning | Screencast |

6 | Sara Hahner | 26 | Rademacher Complexity | Screencast |

7 | Zhuofan Liu | 30 & 31 | Compression Bounds & PAC-Bayes | Screencast |

X | All | All | Collected Errata |