Module Handbook

  • Dynamischer Default-Fachbereich geändert auf INF

Module INF-71-57-M-6

Very Deep Learning - Recent Methods and Technologies (M, 4.0 LP)

Module Identification

Module Number Module Name CP (Effort)
INF-71-57-M-6 Very Deep Learning - Recent Methods and Technologies 4.0 CP (120 h)

Basedata

CP, Effort 4.0 CP = 120 h
Position of the semester 1 Sem. in WiSe
Level [6] Master (General)
Language [EN] English
Module Manager
Lecturers
Area of study [INF-KI] Intelligent Systems
Reference course of study [INF-88.79-SG] M.Sc. Computer Science
Livecycle-State [NORM] Active

Notice

The lecture wants to dive into very deep learning methods, i.e., the latest state-of-the-art. Therefore it is important that you generally know about Machine Learning and Neural Networks. We suggest having a look at the first online lectures of CS231n: https://www.youtube.com/watch?v=g-PvXUjD6qg&list=PLlJy-eBtNFt6EuMxFYRiNRS07MCWN5UIA Especially the content of Lectures 1-5 should be known by every attendee.

Also we pose already now the first exercise, it would be good if you start as early as possible to play around with the standard ML toolkits:

At least one of them is mandatory, but for the best learning outcome we recommend to install all three of them.

Courses

Type/SWS Course Number Title Choice in
Module-Part
Presence-Time /
Self-Study
SL SL is
required for exa.
PL CP Sem.
2V+1U INF-71-57-K-6
Very Deep Learning - Recent Methods and Technologies
P 42 h 78 h
U-Schein
- PL1 4.0 WiSe
  • About [INF-71-57-K-6]: Title: "Very Deep Learning - Recent Methods and Technologies"; Presence-Time: 42 h; Self-Study: 78 h
  • About [INF-71-57-K-6]: The study achievement "[U-Schein] proof of successful participation in the exercise classes (ungraded)" must be obtained.

Examination achievement PL1

  • Form of examination: written exam (Klausur) (60-180 Min.)
  • Examination Frequency: each semester
  • Examination number: 67157 ("Very Deep Learning - Recent Methods and Technologies")

Evaluation of grades

The grade of the module examination is also the module grade.


Contents

In this lecture the most recent advances of deep learning will be presented.

The intended schedule is:

  • Introduction, Motivation
  • Advanced Convolutional Networks (ConvNet, AlexNet, GoogLeNet)
  • SqueezeNet
  • Extended Recurrent Neural Networks (LSTM, MD-LSTM, Dynamic Cortex Memories)
  • Spiking Neural Networks
  • Reinforcement Learning (Policy and Value Networks)
  • Bleeding-Edge Architectures (depending on the most recent publications in Deep Learning).

Competencies / intended learning achievements

Expected outcomes:
  • Understanding and Implementing advanced deep learning methods
  • Solving difficult tasks in Pattern Recognition, Data Science, and Big Data Analytics

Literature

Libaries:

Requirements for attendance of the module (informal)

Knowledge of Neural Networks, MLP, Backpropagation

Recurrent Neural Networks

Strong knowledge in Linear Algebra and theoretical computer science

Python programming skills

Requirements for attendance of the module (formal)

None

References to Module / Module Number [INF-71-57-M-6]

Course of Study Section Choice/Obligation
[EIT-88.A20-SG#2021] M.Sc. European Master in Embedded Computing Systems (EMECS) [2021] [Free Elective Area] Elective Subjects [W] Elective Module
[EIT-88.?-SG#2021] M.Sc. Embedded Computing Systems (ESY) [2021] [Free Elective Area] Elective Subjects [W] Elective Module
[INF-88.79-SG] M.Sc. Computer Science [Specialisation] Specialization 1 [WP] Compulsory Elective
Module-Pool Name
[INF-SIAK-DT-AI-MPOOL-6] SIAK Certificate "Digital Transformation" - Modules INF "Artificial Intelligence"