Master degree in Materials Science

Machine Learning and its Application to Chemistry and Materials Science

Machine Learning and its Application to Chemistry and Materials Science

Academic year 2021/2022

Sommario del corso

• Course objectives
• Results of learning outcomes
• Course delivery
• Learning assessment methods
• Program
• Suggested textbooks and readings
• Teaching tools

Course objectives

The course aims at providing the student with basic concepts of machine learning and at discussing its possible applications in the fields of chemistry and materials science. In particular, the student will learn about the main methods and algorithms of supervised and unsupervised learning (both for regression and classification purposes). Hands-on sessions will provide the student with practical guidelines on how to actually perform machine learning experiments for real case studies (using Python-based tools).

Results of learning outcomes

• Basic concepts of statistics;
• Familiarity with the “jargon” of machine learning;
• Knowledge of the main machine learning approaches for supervised learning (regression, discriminant analysis, decision trees, neural networks, support vector machines);
• Knowledge of the main machine learning approaches for unsupervised learning (principal components analysis; clustering; self-organizing maps);
• Designing a machine learning experiment for specific applications in chemistry and materials science (data, features  and algorithm selection; validation);
• Knowledge of available tools related to machine learning from Python.

Course delivery

The Course consists in frontal lessons, systematically complemented by hands-on sessions, where all concepts are applied by use of available Python-based tools after being formally presented and discussed. [total of 32 hours]

All classes will take place face-to-face (in Aula D1, at the 5-th floor of the building in via Giuria 9) and simultaneously streamed on-line via WebEx on the personal rooms of the teachers:

https://unito.webex.com/meet/alessandro.erba (Lessons 1-4)

https://unito.webex.com/meet/eugenio.alladio (Lessons 5-12)

https://unito.webex.com/meet/lorenzo.mino (Lessons 13-16)

The Course has a strong hands-on character so all students will need to have either a laptop (for face-to-face attendance) or a PC (for on-line attendance).

All students are kindly asked to register to the Course using the relevsant item in the menu bar at the bottom of this page, in order to receive relevant communications about the Course and what needs to be installed on your computer. 

Learning assessment methods

The exam consists in:

1) Preparation of a technical report (to be prepared with the Python tools for Machine Learning discussed throughout the Course) on a specific exercise to be individually assigned to each student by the teachers. The report has to be completed and sent to the teachers 2 days before the actual date of the exam;

2) Oral examination discussing the technical report and other general topics covered in the Course. 

Program

• Introduction (what is machine learning?; Big data in the context of machine learning; general concepts: classification, regression, supervised, unsupervised and reinforcement learning)
• Design and analysis of machine learning experiments (features selection; data pre-processing; model selection; validation and cross-validation; bias and variance; overfitting and underfitting;)
• Supervised learning (linear and multivariate regression; linear and quadratic discriminant analysis; nearest-neighbour approaches; decision trees; elements of neural networks and deep learning; support vector machines)
• Unsupervised learning (principal components analysis; clustering; self-organizing maps)
• Elements of univariate and multivariate statistics (probability distribution functions, maximum likelihood estimation, parametric vs. non-parametric models; hypothesis tests)
• Application of selected algorithms to case studies in chemistry and materials science (with Python)

Lecture notes and scripts by the teachers

Class schedule

• Teaching materials
• Exams and finals
• Go to Moodle
• Visit the forums