MACHINE LEARNING FOR ENVIRONMENTAL APPLICATIONS

Academic year
2022/2023 Syllabus of previous years
Official course title
MACHINE LEARNING FOR ENVIRONMENTAL APPLICATIONS
Course code
PHD153 (AF:411628 AR:222616)
Modality
On campus classes
ECTS credits
6
Degree level
Corso di Dottorato (D.M.45)
Educational sector code
INF/01
Period
2nd Semester
Course year
1
Where
VENEZIA
Moodle
Go to Moodle page
This course provides an introduction to the principles, techniques and applications of machine learning and artificial intelligence. The goal is to develop in students the critical skills needed to wisely choose and implement a data analysis solution based on machine learning techniques.
1. Knowledge and understanding
1.1. acquire the main models for automatic knowledge representation;
1.2. acquire the main models of supervised learning; and
1.3. acquire the main models of unsupervised learning

2. Ability to apply knowledge and understanding
2.1. know how to apply the models studied to real problems;
2.2. know how to critically evaluate the performance and behavior of a model applied to a concrete problem;

3. Judgment skills
3.1. know how to understand which features of various machine learning models best fit a given problem;
3.2. know how to critically evaluate the theoretical characteristics of proposed models;

4. Communication skills
4.1. Know how to communicate the results of an experiment using appropriate terminology;

5. Learning skills
5.1. Know how to critically consult reference texts and their bibliography.
Computational thinking, calculus, linear algebra, statistics, python programming
1. Introduction
1.1 What is learning?
1.2 What is and why machine learning?
1.3 Types of machine learning
1.4 ML in Environmental Science
1.5 ML Pipeline

2. Python and Colab - Recap

3. Data Preprocessing
3.1 Feature normalization/scaling
3.2 Data Imputation
3.3 Feature Selection/Reduction
3.4 Data visualization
3.5 Outlier removal

4. Supervised Learning
4.1 Training a model: dataset splits
4.2 Overfitting/Underfitting problem
4.3 k-NN, SVM, Decision Tree/Random Forest

5. Unsupervised Learning
5.1 k-Means
5.2 Hierarchical Clustering
5.3 DBSCAN
5.4 Spectral Clustering

6. Semi-Supervised Learning

7. AI/Deep Learning
7.1 Optimization, backpropagation, losses
7.2 Multi-Layer Perceptron
7.3 AutoEncoders
7.4 Convolutional Neural Networks
7.5 RNN/LSTM

8. Abnormality Detection (1h)
8.1 Isolation Forest
8.2 One-Class SVM

9. Environmental Science Applications
Written exam plus, at the student's choice, one of the following options:
- an oral presentation of a couple of scientific papers
- development of a data analysis project to support the student's research activity
Frontal/Online lecture, slides and scientific papers
English
written and oral
Definitive programme.
Last update of the programme: 23/05/2022