Advanced Deep Learning for Robotics

SS2019 ADLR IN2349

Content

1. Introduction, Recap ML & DL

2. Advanced Network Architectures

3. Hyperparameter & Architecture Search

4. Adversarial Attacks & Defenses

5. Bayesian Deep Learning

6. Generative Models: VAEs & GANs

7. Data Efficient Learning: Transfer, Semi-Supervised & Active Learning

8. Deep Reinforcement Learning I

9. Deep Reinforcement Learning II

10. Deep Reinforcement Learning III

Lecture slides will be added shortly before the lecture. The password for the slides is communicated in the lecture.

Materials

• I. Goodfellow, Y. Bengio and A. Courville. Deep Learning. MIT Press, 2016.

• Christopher M. Bishop. Pattern Recognition and Machine Learning. Springer, 2006.

• Kevin Murphy. “Machine Learning: A Probabilistic Perspective”, MIT Press 2012


• (R. Sutton and A. Barton. Reinforcement Learning: An Introduction, MIT Press 2018)

References

References cited in the slides.

Exam

• Hints for the exam will be given in one of the last lectures (look in the slides)

• To get an idea of what style of questions to expect, you can find here the exam from an introductory (!!!) lecture on Deep Learning in Robotics I once gave. So, the content of the exam of this year’s advanced (!!!) lecture will differ, but the style will be similar.