Target audience: software engineers, specialists who have solid basic knowledge of the Python programming language.

Training program:

Section 1. Basic theory of neural networks (8 hours)
1.1 Biological and mathematical model of a neuron. From neuron to neural network
1.2 Classifier. Classifier training. Adjustment of weight coefficients. Neural network training
1.3 Handling error propagation with a large number of input nodes and layers. Description of the inverse error distribution. Gradient Descent
1.4 Example of updating neural network weights

Section 2. Creating a neural network in Python (52 hours)
2.1 Building a neural network using the Keras and TensorFlow libraries
2.2 Binary classification
2.3 Training a neural network to detect credit card fraud
2.4 Multi-class classification
2.5 Network training accuracy. Saving and Loading a Model
2.6 Convolutional neural networks
2.7 Face detection and recognition
2.8 Recurrent neural networks
2.9 Final project

Preliminary class schedule:

• June 17 (Wed) - 9.00 - 16.00;
• June 18 (Thursday) - 9.00 - 16.00;
• June 19 (Fri) - 9.00 - 16.00;
• June 22 (Mon) - 9.00 - 16.00;
• June 23 (Tue) - 9.00 - 16.00;
• June 24 (Wed) - 9.00 - 16.00;
• June 25 (Thursday) - 9.00 - 16.00;
• June 26 (Fri) - 9.00 - 12.00, TEST