LIMONGELLI CARLA
SANSONETTI GIUSEPPE
Obiettivi
The goal is to present the fundamental models, methods and techniques of some relevant areas of Artificial Intelligence, with particular reference to heuristic search and Machine Learning, and to use them as tools for the development of innovative technologies. As for Machine Learning, the course will allow students to learn the main methods and algorithms typical of the discipline (supervised, unsupervised and with reinforcement). The lessons and practical exercises carried out during the course will allow the student to acquire analytical and problem solving skills on various domains of interest for the discipline.Programma
LIMONGELLI CARLA
anguage:- Development environments; Jupiter Notebook.
- Python foundations. Data structures in Pyhton.
- Python libraries: NumPy, Pandas, matplotlib, ScikitLearn.
4. Machine Learning:
- Regression (simple linear, multiple).
- Classification (Logistic Regression, Decision Trees, Naïve Bayes).
- Clustering.
- Artificial Neural Networks.
- Reinforcement Learning.
- Introduction to Deep Learning.
- Case studies.
SANSONETTI GIUSEPPE
1. Introduction:- Intelligent Agents.
- AI as "Representation and Search".
2. Problem Solving:
- Uninformed search (breadth-first search, uniform-cost search, depth-first search, Iterative deepening search).
- Heuristic search (Best First search, A *, Heuristic Functions).
- Approximate algorithms (Hill Climbing, Simulated Annealing, etc.)
- Adversarial Search and Games (MiniMax, Alfa-Beta Pruning).
- Introduction to Evolutionary Computation.
3. Introduction to the Python language:
- Development environments; Jupiter Notebook.
- Python foundations. Data structures in Pyhton.
- Python libraries: NumPy, Pandas, matplotlib, ScikitLearn.
4. Machine Learning:
- Regression (simple linear, multiple).
- Classification (Logistic Regression, Decision Trees, Naïve Bayes).
- Clustering.
- Artificial Neural Networks.
- Reinforcement Learning.
- Introduction to Deep Learning.
- Case studies.
Prerequisiti
LIMONGELLI CARLA
Foundation of computer science.SANSONETTI GIUSEPPE
Computer Science FoundationsTesti Adottati
Lecture slides.Bibliografia Di Riferimento
Stuart J. Russell, and Peter Norvig. 2021. Artificial intelligence : a modern approach (4th ed.). Pearson Education, Inc., USA.Modalità Erogazione
In-person classes and in-class labs.Modalità Frequenza
Attendance is not compulsory, but it is strongly recommended.Modalità Valutazione
Written exam and practical laboratory test.Esempi Domande
Given the following **State Space**:**States:** S0, S1, S2, S3, S4, S5.
**Initial state:** S0.
**Goal test:** S5.
**Operator costs:** a: cost 2; b: cost 3; c: cost 4, such that:
- a(S0) = S2; a(S1) = S3; a(S3) = S5;
- b(S2) = S4; b(S1) = S4;
- c(S0) = S1; c(S1) = S2; c(S3) = S4.
Given the following heuristic function h:
h(S0) = 7; h(S1) = 6; h(S2) = 5; h(S3) = 3; h(S4) = 4; h(S5) = 0.
**(7 points)** Execute the A* algorithm using **graph search** until termination. Draw the search tree and, at each step, report the contents of the frontier, the node selected for expansion, and the nodes generated.
**(1 point)** Report the solution found by the algorithm, stating whether or not it is optimal and whether the heuristic function h is admissible. Justify your answers.
---
### EXERCISE 4 (6 POINTS)
Define the **Precision** and **Recall** metrics, explaining their meaning and their role in evaluating Machine Learning systems.
Introduce the concept of a **Confusion Matrix**, describing its structure and showing how Precision and Recall can be calculated from it.
Provide at least one concrete example illustrating the calculation and interpretation of both metrics.
Finally, describe **Precision-Recall Curves**, explaining how they are obtained and graphically representing their typical behaviour.
