Unit ARTIFICIAL INTELLIGENT SYSTEMS

INTELLIGENT APPLICATION DEVELOPMENT

Code	A002039
CFU	6
Teacher	Stefano Marcugini
Teachers	Stefano Marcugini
Hours	42 ore - Stefano Marcugini
Learning activities	Caratterizzante
Area	Discipline informatiche
Sector	INFO-01/A
Type of study-unit	Obbligatorio (Required)
Language of instruction	English
Contents	Functional programming paradigm. Ocaml language. Recursion. Pattern matching. Lists. Trees. Backtracking. Graphs. Search algorithms. Elements of lambda-calculus. Implementation of a parser.
Reference texts	M. Cialdea Mayer, C. Limongelli. Introduzione alla Programmazione Funzionale. Esculapio. http://caml.inria.fr/ (to download programming environment and English documentation)
Educational objectives	Understanding the concepts of functional programming. Ability to build applications. Ability to develop complex data stuctures. Ability to develop intelligent applications.
Prerequisites	None
Teaching methods	Lectures, laboratory exercises
Other information	Website: www.unistudium.unipg.it For the exam schedule, see: https://www.dmi.unipg.it/didattica/corsi-di-studio-in-informatica/informatica-magistrale/calendario-esami
Learning verification modality	Final project and oral exam. The final project is designed to test the ability to correctly apply the theoretical knowledge and understanding of the issues proposed. The oral exam is a discussion lasting about 30 minutes designed to ascertain the level of knowledge and understanding about the theoretical contents of the course reached by the student. Also the oral exam will test the ability of communication of the student and the ability of autonomous organization of the speech. At the request of the student the exam may be taken also in English.
Extended program	Functional programming paradigm. Ocaml language. Recursion. Pattern matching. Lists. Trees. Backtracking. Graphs. Search algorithms. Depth-first search and breadth-first search, euristich search. Branch and bound, A* algorithm. Elements of lambda-calculus. Implementation of a parser.
Obiettivi Agenda 2030 per lo sviluppo sostenibile

INTELLIGENT MODELS

Code	A002038
CFU	6
Teacher	Valentina Franzoni
Teachers	Valentina Franzoni
Hours	42 ore - Valentina Franzoni
Learning activities	Caratterizzante
Area	Discipline informatiche
Sector	INFO-01/A
Type of study-unit	Obbligatorio (Required)
Language of instruction	English
Contents	0 Introduction to AI 1 Agent models 2 State space search, planning 3 Multiagent systems and models 4 Reinforcement learning 5 Simulation of complex systems 6 Social Network Analysis
Reference texts	Testi di riferimento: Appunti del docente disponibili sulla piattaforma www.unistudium.unipg.it Artificial Intelligence: A Modern Approach, Stuart Russell and Peter Norvig Pearson, last edition Network Science Albert Lazlo Barabasi (disponibile online) http://networksciencebook.com/ Reinforcement Learning: An Introduction, Richard S. Sutton and Andrew G. Barto, Second Edition, MIT Press, 2018 (disponibile online http://incompleteideas.net/book/RLbook2018.pdf ) Lectures notes made available on www.unistudium.unipg.it
Educational objectives	Expected learning outcomes. Knowledge oriented goals: Knowledge of the main technique for modeling AI agent based problem domains Knowledge of the main techniques for stata space search, uninformed, informed, local search based, automated reasoning and inference based, automated planning. Reinforcement learning and policy optimization Knowledge of modeling techniques for AI application domains characterized by complex networks Ability oriented goals: Ability to use the acquird knowledge to model, design and implement solutions to real application problems characterized by artificial agents and/or complex networks
Prerequisites	Prerequisites for a fruitful acquisition of the contents of this course can be summarized in knowledge of computer science fundamentale, including algorithms, basic concepts of computational complexity, language grammars,basic concepts of logics, data structures, databases and concurrent and distributed system.
Teaching methods	In class lectures. Lab sessions. Discussion of case studies with the class. Continuous assessment and assignments during the semester. Oral and practical final exam (project). Materials, details on continuous assessmnt will be updated on http://www.unistudium.unipg.it, students are required to know the syllabus and follow the techaing updates published on Unistudium.
Other information	Elearning platform for student-lecturer interaction and online forum communication, upload of material and assignments, http://www.unistudium.unipg.it
Learning verification modality	The exam consists of an oral exam covering the entire syllabus and the presentation of a project (individual or group) to be presented orally to the instructor. For students who regularly attend classes, the oral exam—or parts of it—may be replaced by ongoing assessments (exemptions), leaving the project presentation as the final assessment. In-course assessments are designed during the course based on the class’s level and may vary in number, from 2 (a class with a good level, capable of interacting with the instructor and any tutors and of working independently) to 4 (a class with a lower level, which needs to be guided at every step). The oral exam covers the main topics addressed during the course. The exercises given to candidates consist of theoretical and formal questions on the course topics and the solution of problems involving the analysis and modeling of real-world application domains using the artificial intelligence techniques learned. The project involves developing a software project, agreed upon with the instructor, on applications of artificial intelligence to real-world problems by applying the techniques presented in the course in greater depth. The project requires a final report and a presentation by the participants, which form an integral part of the evaluation. The purpose of the project is to acquire and demonstrate experience in applying the knowledge gained to a real or simulated problem. For information on support services for students with disabilities and/or learning disabilities, visit http://www.unipg.it/disabilita-e-dsa
Extended program	0-Introduction to AI Historical, current, and future perspectives on AI and machine learning 1-Agent models, autonomous agents Reflexive, state-based, inference-based, and learning agents, 2-State space search, planning State-space search models applied to uninformed and informed search, local search, adversarial search, automated planning, logic-based action representation, PDDL 3-Multiagent systems and models Cellular automata, strategy-driven agents, behavioral agents, emerging collective behaviors 4-Reinforcement learning Markov decision processes, policy evaluation, Q-learning, integrating planning and learning, IRL, RLHU 5-Simulation of complex systems Basics of complex AI systems, NetLogo 6-Social Network Analysis Graph theory, random and scale-free networks, Barabási-Albert model, metrics, communities, knowledge networks, social networks, information diffusion, network agents, link prediction
Obiettivi Agenda 2030 per lo sviluppo sostenibile	Quality Education Industry , Innovation, Infrastrucures Sustainable cities and communities