Course content:
– Machine Learning Refresh
o Supervised Learning vs. Unsupervised Learning
o Traditional Machine Learning vs. End-to-End Learning
– Fundamentals of Neuronal Networks:
o Rosenblatt Perceptron and Neurons
o Network Structure (feed-forward, recurrent), matrix notation, forward evaluation
– Training as optimization
o Loss and Error functions
o Backpropagation
o SGD and other optimizer
– Activation functions and topologies
o Convolutional neural networks
o Generative Adversarial Networks
o Long short-term memory networks
o Special layer types (inception, resnet)
o Embeddings
o Attention Mechanis & Transformer
– Applications to real-world problems:
o Acoustic keyword recognition (audio/speech processing)
o Sentiment analysis (text processing)
o Digit recognition (image processing)
o Tiny Image Recognition (image processing)
o Face Detection and Tracking (image/video processing)
o Stock market prediction (time series prediction)
– Training on large data sets (Hardware, GPU)
– Trustworthy AI
Structure:
The course is a theoretical content in the morning and practical exercises in the afternoon in form of lab Jupyter notebook programming.