Politecnico di Torino
Politecnico di Torino
   
Login  
it
Politecnico di Torino
Academic Year 2016/17
01BITOV
Artificial intelligence
Master of science-level of the Bologna process in Computer Engineering - Torino
Teacher Status SSD Les Ex Lab Tut Years teaching
Piccolo Elio ORARIO RICEVIMENTO     50 10 0 0 7
SSD CFU Activities Area context
ING-INF/05 6 D - A scelta dello studente A scelta dello studente
Subject fundamentals
The course is taught in Italian.
This course is optional for the Master of Science in Computer Engineering, and it is held at the second semester of the second year.
This course illustrates Artificial Intelligence problems and the approaches for solving them. Its main topics are the basic models for intelligent behavior and their simulation, knowledge representation, the limits to which intelligence is described by the rule evaluations, by inference and deduction. Another important topic is learning systems architectures, and their representation of the external word.
Expected learning outcomes
Knowledge of problem solving techniques, skill to analyze and design programs solving complex problems (planning, problems with constraints, games, etc.).
Knowledge of propositional and first order logic; ability to model different aspects of the world (demonstration, planning, monitoring circuits, etc..) by means of logic.
Knowledge of methods for knowledge representation, skill to design knowledge-based systems (Semantic Web, etc.)
Knowledge of methods to deal with systems in the presence of uncertainty, ability to design in complex reasoning systems (expert systems, etc.)
Knowledge of basic methods for pattern recognition, ability to implement classification systems.
Knowledge of the neural network models and their training techniques; ability to apply neural networks for categorization and pattern recognition problem solving.
Prerequisites / Assumed knowledge
None. Knowledge of Formal Languages, and Data structure and Algorithms is useful, but not mandatory.
Contents
Strategies for problem solving (14 hours):
- Solutions in the state space
- Solution for decomposition into sub-problems
- Research in breadth , depth, and using heuristics
Logic ( 12 hours) :
- The propositional logic
- The first-order logic
- The non-monotonic logic (outline)
- Decision-making
Knowledge representation (12 hours):
- Semantic networks
- The production rules
- The frames
- The hybrid approaches
- Comparisons in terms of expressiveness , deductive power , applicability
Models of reasoning and learning : uncertainty , Bayesian inference , belief (4 hours)
Knowledge-based Systems (4):
- Expert systems : problems and classifications, with particular regard to the application of the same in the areas of technical, engineering ;
- Automatic learning ; user interface in the context of knowledge-based systems (outline) ;
Recognition of configurations ( pattern recognition ) (6 hours):
- Pre-processing and features extraction
- Decision Functions
- Methods of classification
- Comparison using dynamic programming
Architectures that mimic biological systems : neural networks, connectionism , distributed-memory sparse (8 hours)rem - Hidden Markov models
Delivery modes
Different projects will be developed related to the methods introduced in the lectures, in particular: techniques for rules evaluation, expert systems on restricted domains and shell of expert systems, language recognition systems, neural networks, intelligent games, image and speech recognition.
The student has to develop a project referring to a topic introduced in class including some experimental tests.
Texts, readings, handouts and other learning resources
Reference books:
Stuart J. Russell, Peter Norvig, "Artificial Intelligence. A Modern Approach ", Vol. 1 e 2, Pearson, Milano.
E. Rich, "Artificial Intelligence", McGraw Hill, Milano.
N.J. Nilsson, "Problem-Solving Methods in Artificial Intelligence ", McGraw-Hill, New York, 1971

Auxiliary books:
Nils J. Nilsson, " Artificial Intelligence: a New Synthesis", Morgan Kaufman Publishers, Inc.
I. Bratko, "Prolog Programming for Artificial Intelligence", Addison-Wesley/Pearson Education, 2001.

The slides are available on the course's WEB site.
Assessment and grading criteria
The exam is written, with a possible oral examination. The student can select and develop a subject of the course performing an assignment and presenting a short written dissertation.

Programma definitivo per l'A.A.2018/19
Back



© Politecnico di Torino
Corso Duca degli Abruzzi, 24 - 10129 Torino, ITALY
WCAG 2.0 (Level AA)
Contatti