Artificial Intelligence Course Presentation Summary Artificial - - PDF document

Artificial Intelligence

Artificial Intelligence

Course Presentation

Artificial Intelligence

Summary

Motivations Course Plan Resources Exam Methods

Artificial Intelligence

Motivations

Artificial Intelligence: Machines that think and act like humans do Voight-Kampff test in blade-runner

Artificial Intelligence

Motivations

Artificial Intelligence: Machines that solve complex problems Google Self Driving car

Artificial Intelligence

Related areas

AI highly interdisciplinary Probability and Statistics Robotics Logics Algorithms Game Theory Pattern Recognition and Machine Learning

Artificial Intelligence

Practical applications: Overview

Surveillance Environmental monitoring Search and Rescue operations Energy management Service Robots Games, entertainment and education Computer Vision Medical Diagnosis Hardware/Software Verification ...

Artificial Intelligence

Service Robots/Entertainment: Cooperative Foraging

Decide who is in the best position to execute a task

Artificial Intelligence

Surveillance and Monitoring: mobile sensor exploration

A group of sensors cooperatively plans for most informative paths

Artificial Intelligence

Surveillance and Monitoring: precisione agriculture

Analyse data from greenhouse sensor network to maximize crop yield and minimize infection (Post-doc: Alberto Castellini, Project: EXPO-AGRI)

Artificial Intelligence

Surveillance and Monitoring: Multi-Robot Patrolling

Allocate visit locations to a group of robots

Artificial Intelligence

Security: Active Malware Analysis

Use Reinforcememnt Learning to analyse malware behaviors (PhD: Riccardo Sartea)

Artificial Intelligence

Ride-Sharing: coalition formation

Form groups of riders to minimize fuel consumption (Post-Doc: Filippo Bistaffa)

Artificial Intelligence

Environmental Survey: Water Monitoring

Intelligent drones to monitor water quality

Artificial Intelligence

Water Monitoring: High level control for the drones

Human interaction with team oriented plans (PhD student: Masoume Raeissi)

Artificial Intelligence

Water Monitoring: Planning informative paths

Active learning to devise informative paths for classification (PhD student: Lorenzo Bottarelli)

Artificial Intelligence

Water Monitoring: perception for autonomous behaviors

Use computer vision to detect relevant features and situations (Researcher: Domenico Bloisi)

Artificial Intelligence

Course Plan I

Problem Solving: Search (about 4 Lessons)

Uninformed search (Breadth first, Depth First, Iterative Deepening, etc.) Informed Search (A*, Heuristics, Local Search and Optimization)

Constraint Processing (CSP , COP) (about 4 lessons)

Contraint Satisfaction Problems, Constraint Network and Graphical models Basic techniques for CSP (Consistency enforcing, Backtracking, Local Search) Tree-Decomposition (Dynamic Programming) Constraint Optimisation Problems

Artificial Intelligence

Course Plan II

Multi-Agent Systems (about 2 lessons)

Distributed COPs Reaching agreement

Prova parziale (approx. end of April) Adversarial Search (1 lesson) Plan representation and monitoring (1 lesson) Logic and Agents (about 2 lessons)

Logical Agents Background on Logic (propositional, FOL) Inference (DPLL, Resolution)

Probabilistic Reasoning (about 5 lessons)

background on Probability Bayesian Network Inference (complete and approximate) Markov Decision Processes and Reinforcement Learning

Artificial Intelligence

Resources

Text Books

Artificial Intelligence: a modern approach 2nd Editon Russel and Norvig (English edition) Constraint Processing R. Dechter

Other Material

Scientific Papers, Slides, etc. Will be available on web site

Web Page link http://profs.sci.univr.it/ farinelli/courses/ia/ia.html

Artificial Intelligence

Exam modalities

Single-test mode

Single written test at the exam day

Partial test mode: Two tests C1 + [C2 or P]

C1 and C2: solve simple exercises/describe techniques studied during the course P:

project to be developed at home (see below)

• nly to the exams right at the end of the class (Summer

Session) partial written test C1: half-way through the course C2: at the end of the course. project (P) can be done in collaboration with another person Final grade: 50%C + 50%[C1 or P]

Artificial Intelligence

Projects

Project

Instructor will propose a set of projects Students can: choose among the set of proposed projects or propose other projects Projects proposed by students must be validated by the instructor Projects usually involve a programming part (in the language most appropriate for the project) Students must hand to the instructor a report of the project and developed code. Have a look at past projects on the course web site