Students are expected to have sufficient background in calculus, linear algebra, programming, and quantum computing.

This course introduces the fundamental concepts of neural networks and the backpropagation algorithm. We shall explore how this architecture can be applied in the design and automatic learning of quantum circuits for certain tasks. We will present extensions of classical to quantum neural networks and analyze their performance for classical and quantum learning problems. Students attending this course will: - Understand the theoretical basis and the most common architectures of neural networks. - Study the different architectures of hybrid-quantum models and quantum neural networks. - Know and use automatic differentiation software packages for training quantum learning models. - Apply a quantum neural network for simple learning tasks with classical and quantum training data.

Unit 1. Introduction to neural networks 1.1. Perceptron, layers, and backpropagation algorithm 1.2. Deep architectures and methods for correlated data Unit 2. Classical-quantum hybrid models 2.1. Parametric quantum circuits 2.2. Training datasets and loss functions 2.3. Learning quantum algorithms Unit 3. Quantum neural networks (QNN) 3.1. Quantum models of a perceptron 3.2. QNN for classical learning tasks 3.3. Quantum learning tasks

- Theoretical sessions presenting the fundamentals of neural networks and backpropagation. - Practical sessions on automatic differentiation software packages for model training. - Practical labs implementing and training classical-quantum hybrid models and QNNs. - Tutorial sessions. - Student individual and team work.