Vinaora Nivo Slider 3.x

Elements of Artificial Intelligence

CFU: 6

Prerequisites

None.

Preliminary Courses

None.

Learning Goals 

The course aims at providing the basic methodologies and techniques to understand and address issues related to Artificial Intelligence.

Students will acquire the theoretical background related to intelligent agents, their interaction, problem-solving, search strategies and adversarial search. They will learn the methods and techniques in the domain of game theory, which include optimal, imperfect real-time decisions, games with random elements, and state-of-the-art of game programs.

Students will acquire the basics of first-order logic, inference, and deduction, as wells as they will master methods and techniques of logic programming with ProLog. They will be able to model uncertain knowledge and reasoning in order to act in uncertainty. Finally, the course will introduce basic concepts behind probabilistic reasoning and machine learning.

Expected Learning Outcomes 

Knowledge and understanding

The course aims to provide students with the knowledge needed to understand and analyze problem solutions based on Artificial Intelligence techniques.

Tools to master both the theory and the methodologies for problem solving will be provided; in particular, search strategies, adversarial search and logic programming will be considered. Basic concepts behind probabilistic reasoning and machine learning will be also introduced.

Applying knowledge and understanding

The course focuses on conveying the skills and methodological, as well as operational tools, which are the bases to apply Artificial Intelligence knowledge. Lessons promote the ability to apply the acquired methodological tools to implement solutions based on Artificial Intelligence techniques. The proposed techniques and models will be applied to specialized domains.

Course Content - Syllabus

Part I: Introduction to Artificial Intelligence

Intelligent Agents: Agents and environments, the concept of rationality, the nature of environments, the structure of agents

 

Part II: Problem Solving

Solving problems with search: Problem solving agents, Example problems, Looking for solutions, Uninformed search strategies, Breadth search, Uniform cost search, Depth search, Limited depth search

Iterative in-depth search, Two-way search, Comparison of uninformed search strategies, Avoiding repetition in states, Searching with partial information.

Informed search: Informed search strategies or heuristics, Best-first greedy or "greedy" search, A* search, Heuristic search with limited memory, Local search algorithms and optimization problems, Hill-climbing search, Simulated annealing, Local-beam search, Genetic algorithms.

Searching with opponents: Games, Optimal decisions in games, The minimax algorithm, Alpha-beta pruning, Imperfect real-time decisions, Games that include random elements, The state of the art in game programs.

 

Part III: Knowledge and Reasoning

Logical agents: Knowledge-based agents, The world of wumpus, Logic, Propositional calculus, Patterns of reasoning in propositional calculus, Forward and backward concatenation.

First-order logic: Syntax and semantics of first-order logic, Using first-order logic.

Inference in first-order logic: Propositional inference and first-order inference, Unification

Forward Concatenation, Backward Concatenation, Logic Programming, Prolog, Lists in Prolog, Extra-logic Operators: not, cut, fail

 

Part IV: Uncertain Knowledge and Reasoning

Uncertainty: Acting under uncertainty, Basic notation of probability theory, Inference based on complete joint distributions, Independence, Bayes' rule and its use.

Probabilistic reasoning: Representation of knowledge in an uncertain domain, Semantics of Bayesian networks

Efficient representation of conditional distributions.

 

Part V: Learning

Learning from observations: Forms of learning, Inductive learning, Learning decision trees.

Neural Networks: Definition of Neural Network, Training and Learning, Training Modes, Learning Laws

The perceptron of Rosenblatt, The multilevel perceptron, The theorem of Kolmogorov, Learning Vector Quantization (LVQ) Network, Kohonen Self-Organizing Maps (SOM), Kernel Machines, Support Vector Machines (SVM).

Readings/Bibliography

Recommended textbooks:

    S.J.Russell, P. Norvig, Artificial Intelligence: A Modern Approach, Global Edition, Third Edition, Pearson Education.

Other materials:

   Materials produced and provided by the Teacher

Teaching Methods

The teaching is carried out with lectures (70% of total hours) and laboratory exercises (30% of total hours).

Examination/Evaluation criteria

Exam type

Written and oral and also a project discussion. Questions of the written exam refer to open answers. The project will be proposed in the middle of the course.

Evaluation pattern

The examination aims at verifying the achievement of the formative objectives expected for the teaching activities. It includes a written test and an oral discussion focused on the topics of the course.

We use cookies on our website. Some of them are essential for the operation of the site, while others help us to improve this site and the user experience (tracking cookies). You can decide for yourself whether you want to allow cookies or not. Please note that if you reject them, you may not be able to use all the functionalities of the site.