Vorlesung: Deep Learning for Visual Recognition


  • Dozent(en):
  • Beginn: 19.10.2016
  • Zeiten: Wed. 10:30 - 12:00, LBH / HS III.a
  • Veranstaltungsnummer: MA-INF 2313
  • Studiengang: Master
  • Aufwand: 6 CP
  • Prüfungen: Feb. 21, 13:00 - 15:00, LBH / HS III.a



Neural networks are making a comeback!  Deep learning has taken over the machine learning community by storm, with success both in research and commercially.  Deep learning is applicable over a range of fields such as computer vision, speech recognition, natural language processing, robotics, etc.  This course will introduce the fundamentals of neural networks and then progress to state-of-the-art convolutional and recurrent neural networks as well as their use in applications for visual recognition.  Students will get a chance to learn how to implement and train their own network for visual recognition tasks such as object recognition, image segmentation and caption generation.

No formal pre-requisites.  Students should already be comfortable with concepts in probability theory and optimization and are recommended to have taken at least one course in machine learning or computer vision.  Exercises will be a mix of theory and practical (Python).


The second exams have been graded and should already be available on BASIS.  If you want to look at your marked exam, you can stop by I.81 on Wednesday Apr. 5 from 14:30 - 16:30.  

The second exam will be held on Mar. 30 from 9:00 - 11:00 in LBH III.a.  Please bring photo ID. 

Exams have been graded and will be available on BASIS by Friday Mar. 3.  If you want to look at your marked exam, you can stop by I.81 on Wednesday Mar. 8 from 14:30 - 16:30.  

If you enjoyed the class and would like to do some more hands-on work with deep learning and visual computing in a HiWi position starting June 2017, send me an email along with your CV, transcript and reports for previous seminars and or labs.

The final exam will be held on Feb. 21 from 13:00 - 15:00 in LBH III.a.  Please bring photo ID.

Project presentations will be held during the exercise session on Feb. 1 (14:30 - 16:00), the second half of lecture on Feb. 8 (11:45 - 12:00) and the exercise session. You must submit your source code and your presentation in .pdf format to the assistants before 13:00 if you are presenting on Feb. 1 and before 9:00 if you are presenting on Feb. 8.  

The problem in BASIS has been fixed and you should now be able to register for exams. Please notify Fadime if you still cannot register.

No lecture and no exercise on December 7 due to Dies Academicus!

Instructions for accessing the GPU server for the projects have been posted online under Additional Documents.  Log-in credentials and the link to the scheduling doodle have been sent out via email.  If your group did not receive such an email, please contact Soumajit.  Please try to log on asap; if the account does not work, please contact Soumajit.  GPU scheduling is done on a first-come first serve basis, so sign up at your earliest convenience.

Please note that there will be double lecture on Nov. 30, with the second lecture taking place in the exercise slot.  Course projects will also be announced on this day in the exercise slot.  Please make sure you are present to sign up for a project and a presentation date.  The projects will be allocated on a first come first serve basis.  If you are unable to attend, please email the TAs to make individual arrangements.

First lecture starts on October 19, 2016.  There will be no exercises on the first day.  See you there!

Weitere Dokumente



Übung 1: MLbasics
Übungsblatt  (PDF-Dokument, 157 KB)
Übung 2: Theano
Übungsblatt  (einfaches Textdokument, 30 Bytes)
Übung 3: NNintro
Übungsblatt  (PDF-Dokument, 109 KB)
Übung 4: Optimization
Übungsblatt  (PDF-Dokument, 139 KB)
Übung 5: CNNs
Übungsblatt  (PDF-Dokument, 116 KB)
Übung 6: RNNs
Übungsblatt  (PDF-Dokument, 217 KB)
Übung 7: Autoencoders
Übungsblatt  (PDF-Dokument, 144 KB)