Artificial Neural Networks 2401-CS-21-ANN-s2

Lecture:

1. Introduction  
- Artificial Neuron 
- Activation function 
- How do Artificial Neural Networks (ANN) learn
 - Gradient descent  & Stochastic gradient descent
- Back propagation
- Evaluating and improving ANN

2. Computer Vision (CV):
- What are convolutional Neural Networks (CNN)
- Why CNNs for CV
- CNN layer types (Convolutional Layer, Pooling Layer)
- Case study: Image recognition and classification

3. Natural Language Processing (NLP) 
- The idea behind Recurrent Neural Networks 
- The vanishing gradient problem
 - LSTM

4. Transformers 
- Attention Mechanism

5. Summary and going forward: Multimodal and Multitask learning

Laboratory:
Introduction to ANN and PyTorch
Project 1: Multilayer perceptrons - part 1
Project 1: Multilayer perceptrons - part 2
Project 2: Computer Vision - part 1
Project 2: Computer Vision - part 2
Project 3: NLP - part 1
Project 3: NLP - part 2
Presentation of projects

Efekty uczenia się - wiedza

1: Student knows how to apply basic artificial neural networks to solve real life problems. 2: Student understands the most important deep learning algorithms and methods. 3: has basic knowledge of differences between artificial neural networks: Fully Multilayer Perceptron, Convolutional Neural Network and Recurrent Neural Networks.

Efekty uczenia się - umiejętności

1: Student is able to efficiently communicate deep learning results to the diverse audience 2: Student is able to use modern tools and methods to solve complex real-life problems. 3:Student is capable of creating simple artificial neural network and test its learning processes.

Wymagania wstępne

Basic programming with Python, basics of statistics and probability, basics of linear algebra, basics of mathematical analysis

Koordynatorzy przedmiotu

Bogna Bylicka

Literatura

Fundamental literature:
L. Antiga, E. Stevens, T. Viehmann, Deep Learning with PyTorch, O’Reilly 2020
DataCamp course on DL with PyTrch
Supplementary literature:
J. Brownlee, Deep Learning Mastery with Python, www.machinelearningmastery.com
Internet Courses on DeepLearning, especially on Coursera, Udacity.com and Udemy.com

Więcej informacji

Dodatkowe informacje (np. o kalendarzu rejestracji, prowadzących zajęcia, lokalizacji i terminach zajęć) mogą być dostępne w serwisie USOSweb:

Strona przedmiotu 2401-CS-21-ANN-s2 w USOSweb