Advanced Machine Learning
(MSc)
Lecture Notes (check Index for related concepts)
Topics:
1. Introduction to Deep Learning
2. Neural Networks and Backpropagation
3. Embeddings and Recommender Systems
4. Convolutional Neural Networks for Image Classification
5. Deep Learning for Object Detection and Image Segmentation
6. Recurrent Neural Networks and NLP
7. Sequence to Sequence, Attention and Memory
8. Expressivity, Optimization and Generalization
9. Imbalanced Classification and Metric Learning
10. Unsupervised Deep Learning and Generative models
11. Ensemble Learning
12. Autoencoders and Decoders
13. Visualizing Convolutional Neural Networks
14. Software Packages
Literature:
ˇ W. McKinney: Python for Data Analysis (1 ed.). O'Reilly Media, Inc. 2012.
ˇ D. Conway, J.M. White: Machine Learning for Hackers, O'Reilly Media, Inc., 2012.
ˇ Ludmila I. Kuncheva: Combining Pattern Classifiers: Methods and Algorithms, Second Edition, Wiley, 2014
ˇ I. Goodfellow, Y. Bengio, A. Courville: Deep Learning, MIT Press, 2016.
ˇ Francois Chollet: Deep learning with Python, Manning, November 2017.
Main concepts (for exam):
Neural networks (Activation
functions, Backpropagation,
Stochastic gradient descent, Momentum, Optimizers), Embeddings, Dropout Regularization, Recommender
Systems, Convolutions, CNNs
for Image Classification,
CNN Architectures, Simple Localisation as regression, Detection Algorithms, Fully convolutional Networks, Semantic & Instance Segmentation, Natural language processing (Classification and word representation, Word2Vec, Language
Modelling, Recurrent neural
networks), Encoder-decoder for machine translation,
Attention mechanisms, Towards memory and reasoning, Autoencoders, Generative Adversarial Networks
Lecture notes for
exam
For interested
readers (references from the slides):
Introduction/history
Backpropagation
Optimization in gradient descent
Optimization
in gradient descent 2