Deep learning has brought a strong leap forward in many pattern recognition applications. Optimists even see a fundamental revolution questioning the supremacy of human cognitive powers.
How does deep learning work? Is it indeed the revolution that some announce? What are its limits and how can it be applied for multimedia pattern recognition?
In this lecture, we will discuss both the theoretical basis of deep learning and numerous applications based on data from modalities like audio, image, and video.
19.04.2016: Neural Networks
26.04.2016: Deep Neural Networks
10.05.2016: Theory 1: Universal approximation
24.05.2016: Theory 2: NP hardness of NN training
31.05.2016: Recurrent Neural Networks
07.06.2016: Applications 1: Image Recognition
14.06.2016: Applications 2: Speech Recognition
21.06.2016: Applications 3: Audio and Music Analysis
28.06.2016: Applications 4: Video Analysis
05.07.2016: Limits of Deep Learning
Materials for participants