Programmazione quantistica e Intelligenza Artificiale
Academic year: 2024/25
Semester: 1
CFU: 6
Hours: 45
Teachers
Syllabus
Python and Qiskit Programming (3 CFU)
- Python
- IBM Quantum
- Qiskit
- Advanced functions
Machine Learning for Quantum (3 CFU)
- Introduction to Machine Learning
- What is Machine Learning?
- Categories: Supervised, Unsupervised, Reinforcement Learning
- Typical ML workflow: dataset, preprocessing, training, evaluation
- Supervised Learning fundamentals: input/output, loss functions, overfitting vs. underfitting
- Linear Regression: model, cost function, gradient descent
- Objective: Gain an overview of ML and solve a basic regression task.
- Classification & Evaluation Metrics
- Binary vs. Multiclass Classification
- K-Nearest Neighbors
- Model evaluation: accuracy, precision, recall, F1 score, ROC-AUC
- Train/test split and cross-validation
- Introduction to the Fisher Method (Linear Discriminant Analysis)
- Objective: Learn how to evaluate classification models and solve a classification problem.
- Unsupervised Learning & K-means Clustering
- Introduction to Clustering
- K-means: algorithm, Euclidean distance, choosing number of clusters
- Dimensionality reduction for visualization (PCA, t-SNE)
- Comparison between clustering and classification
- Objective: Understand the logic of unsupervised learning and apply clustering to real-world data.
- Decision Trees & Ensemble Methods
- Decision Trees: structure, entropy, information gain, overfitting
- Introduction to Random Forests
- Overview of Bagging vs. Boosting
- Pros and cons of tree-based models
- Objective: Understand interpretable models and be introduced to ensemble learning techniques.
- Feedforward Neural Networks & Backpropagation
- Perceptron: historical context and basic model
- Multi-Layer Perceptron (MLP) architecture
- Activation functions: sigmoid, ReLU, softmax
- Backpropagation: intuition and structure
- Training and challenges (e.g., vanishing gradient)
- Objective: Grasp how basic neural networks function and implement a simple MLP.
- Autoencoders & Advanced Topics Overview
- What is an autoencoder? (use cases: compression, denoising)
- Architecture: encoder/decoder
- Comparison with PCA
- Practical examples: latent space visualization
- Brief overview of advanced architectures (CNNs, RNNs) to give context
- Objective: Introduce unsupervised neural architectures and prepare the ground for more advanced or quantum-related topics
- A Primer on Reinforcement Learning
- Markov Decision Processes (MDPs)
- Policies, Goals, Rewards, and Returns
- Value Functions
- Policy-based vs. Value-based Methods
- Policy Gradient Methods and the REINFORCE Algorithm
- Q-Learning Algorithm
- Objective: Introduce the fundamental concepts of reinforcement learning, equipping students with a clear understanding of policy- and value-based approaches.
- Reinforcement Learning for Quantum Control
- Population Transfer in a Three-Level Quantum System
- Stimulated Raman Adiabatic Passage (STIRAP) Protocol
- Application of Reinforcement Learning to Solve the Three-Level Population Transfer Problem
- Objective: Learn how to apply reinforcement learning to effectively address quantum control challenges.
- Supervised Learning for Quantum Control and Noise Detection
- Quantum Control of a Qubit via Supervised Learning:
- Problem Setup and Quantum Gates
- Implementing and Evaluating Solutions
- Detection and Classification of Noise Correlations in a Three-Level Quantum System
- Objective: Learn how supervised learning techniques can be applied to quantum control problems and noise characterization.
Bibliography
- A. Géron, Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow (2nd ed.) (O’Reilly Media, 2019). It covers sections 1, 2, 3, 4, 5 and 6. Full guide for machine learning experts and beginners.
- C. M. Bishop, Pattern Recognition and Machine Learning (Springer, 2006). It covers sections 1, 2, 3 and part of the 4. For full immersion in mathematics and statistics.
- I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning (MIT Press, 2016). It covers sections 5 and 6. For brave warriors.
- R. S. Sutton, A. G. Barto, Reinforcement Learning: An Introduction (MIT press, 2018).
- Florian Marquardt, Machine learning and quantum devices, SciPost Physics Lecture Notes 29 (2021).
- L. Giannelli, S. Sgroi, J. Brown, G. S. Paraoanu, M. Paternostro, E. Paladino, and G. Falci, A Tutorial on Optimal Control and Reinforcement Learning Methods for Quantum Technologies, Physics Letters A 434, 128054 (2022).
- S. Mukherjee, D. Penna, F. Cirinnà, M. Paternostro, E. Paladino, G. Falci, and L. Giannelli, Noise Classification in Three-Level Quantum Networks by Machine Learning, Machine Learning: Science and Technology 5, 045049 (2024).