ARTIFICIAL INTELLIGENCE
FALL 2014
Solutions for Assignment 3
Problem 2:
a) We need to evaluate the probability P(, b, c, d, e). Given the Bayesian network we have that:
P(, b, c, d, e) = P()P(b)P(c)P(d|, b)P(e|b, c) =
0.2 0.5 0.8 0.1 0.3 = 0.0024
(4 point
Exam Study Guide: Baynesian Networks
Bayesian Networks: Properties and Exact Inference
Bayesian network is a graph where:
the nodes correspond to the random variables of a problem
directed linkes connect pairs of nodes
each node has a conditional probabil
CS 440/520: INTRODUCTION TO ARTIFICIAL INTELLIGENCE
FALL 2014
Assignment 1
Fast Trajectory Replanning
Deadline: October 5, 11:59pm.
Available points for undergraduate students: 120. Perfect score: 100.
Available points for graduate students: 105. Perfect
CS 440/520: INTRODUCTION TO ARTIFICIAL INTELLIGENCE
FALL 2014
Assignment 5
Inductive Learning, Support Vector Machines, Neural Networks
Deadline: December 10, 11:59pm.
Available points for undergraduate students: 105. Perfect score: 100.
Available points
CS 440/520: INTRODUCTION TO ARTIFICIAL INTELLIGENCE
FALL 2014
Assignment 3
Classical Planning - Probabilistic Reasoning - Bayesian Networks and Exact Inference Approximate Inference - Value of Information
Deadline: November 9, 11:59pm.
Available points fo
SupportVectorMachinesNotes
1.
What is the idea of a margin in the context of support vector
machines and how does it relate to a separating hyperplane? Provide a graphical
example.Whatisthesupportvectorinthesamecontext?
Themargininthecontextofsupportvecto
Study Notes for AI (CS 440/520)
Lecture 9: Classical Planning
Corresponding Book Chapters: 10.1-10.2
Note: These notes provide only a short summary and some highlights of the material covered in the corresponding
lecture based on notes collected from stud
Study Notes for AI (CS 440/520)
Lecture 5: Local Search
Corresponding Book Chapters: 4.1-4.2-4.5
Note: These notes provide only a short summary and some highlights of the material covered in the corresponding
lecture based on notes collected from students
Study Notes for AI (CS 440/520)
Lecture 11: Bayesian Networks and Exact Inference
Corresponding Book Chapters: 14.1-14.2-14.3
Note: These notes provide only a short summary and some highlights of the material covered in the corresponding
lecture based on
Study Notes for AI (CS 440/520)
Lecture 6: Constraint Satisfaction Problems
Corresponding Book Chapters: 6.1-6.2-6.3-6.4
Note: These notes provide only a short summary and some highlights of the material covered in the corresponding
lecture based on notes
Study Notes for AI (CS 440/520)
Lecture 12: Approximate Inference in Bayesian Networks
Corresponding Book Chapters: 14.4-14.5
Note: These notes provide only a short summary and some highlights of the material covered in the corresponding
lecture based on
Study Notes for AI (CS 440/520)
Lecture 10: Intro to Decision and Probability Theory
Corresponding Book Chapters: 13.2-13.3-13.4-13.5-13.6
Note: These notes provide only a short summary and some highlights of the material covered in the corresponding
lect
Markov Decision Processes Review
(Partially Observable)
1. What do we need to define in order to formulate a Markov Decision Process?
In order to formulate a MDP you need to define:
The transition model:
, ,
- probability of going from state
s to state s
Utility Theory
Study Guide
The basic principle of utility theory is that there is a single # which expresses the
desirability of the state. An agent wants to maximize its utility by choosing actions that
give it the maximum utility.
How can we compute the
Dynamic Bayesian Networks: Temporal State Estimation
Markov Assumption current state depends on only a finite fixed number of previous
states. Simplest is first-order Markov process, in which current state depends only on
previous state and not on any ear
ArtificialNeuralNetworksNotes
1.
Giveagraphicalrepresentationofasingleartificialneuron.Writetheoutputof
theneuronasafunctionoftheinput.
Thesimplemodelhas3partstoit.Therestheinputfunction,theactivationfunctionandthe
output. The input function is simply the
Decision Networks and Value of Information
How is a decision network different than a Bayesian network? Describe the purpose of
the additional nodes.
Decision networks vary from Bayesian networks because decision
networks contain additional nodes. They ha
Intro. to Artificial Intelligence
Bayesian Learning
Questions and Answers
1.
What is the idea in Bayesian learning for evaluating how
probable a hypothesis is based on observed data? How can we make
predictions using observed data in the context of Bayesi
CS 440/520: INTRODUCTION TO ARTIFICIAL INTELLIGENCE
FALL 2014
Assignment 2
Local and Classical Search - Adversarial Search Constraint Satisfaction Problems - Logic-based Inference and Satisability
Deadline: October 17, 11:59pm.
Available points for underg
Study Notes for AI (CS 440/520)
Lectures 7 & 8: Logic-based Inference and Satisability
Corresponding Book Chapters:
Note: These notes provide only a short summary and some highlights of the material covered in the corresponding
lecture based on notes coll
Study Notes for AI (CS 440/520)
Lecture 2: Uninformed Search
Corresponding Book Chapters: 3.1-3.2-3.3-3.4
Note: These notes provide only a short summary and some highlights of the material covered in the corresponding
lecture based on notes collected from
Study Notes for AI (CS 440/520)
Lecture 4: Adversarial Search
Corresponding Book Chapters: 5.1-5.2-5.3-5.4-5.5
Note: These notes provide only a short summary and some highlights of the material covered in the corresponding
lecture based on notes collected
Satisability
Suggested Format
Last revision: May 8, 1993
This paper outlines a suggested format for satisability problems. It is not
yet the \oial" DIMACS graph format. If you have omments on this or
other formats or you have information you think should
CS 440: INTRODUCTION TO ARTIFICIAL INTELLIGENCE
SPRING 2017
Assignment 1
Heuristic Search using Information from Many Heuristics
Deadline for the first phase: February 5, 11:59pm
Deadline for the second phase: February 15, 11:59pm
Perfect score: 100 point
Computer Science Department
CS520
IntrotoArtificialIntelligence
Recitation 4
Local search
OptimizationProblems
Instead of considering the whole state space, consider only
the current state
Limits necessary memory; paths not retained
Amenable to large o
Adversarial Search
Game Playing
An AI Favorite
structured task, often a symbol of intelligence
clear definition of success and failure
does not require large amounts of knowledge
(at first glance)
focus on games of perfect information
multiplayer, ch
Temporal probability models
Chapter 15, Sections 13
c
Artificial Intelligence, spring 2013, Peter Ljungl
of; based on AIMA Slides Stuart
Russel and Peter Norvig, 2004
Chapter 15, Sections 13
1
Outline
Time and uncertainty
Inference: filtering, predictio
CS 440: INTRODUCTION TO ARTIFICIAL INTELLIGENCE
SPRING 2017
Assignment 4
Decision Making under Uncertainty and Learning
Deadline: May 1, 11:59pm
Available points: 110 - Perfect score: 100 points.
Assignment Instructions:
Teams: Assignments should be compl
CS 188
Fall 2010
Introduction to
Artificial Intelligence
Section Handout 9 Solutions
Value Of Information
Used Car Purchase
[Adapted from problem 16.11 in Russell & Norvig]
A used car buyer can decide to carry out various tests with various costs (e.g., k