/*
* This code is created for cs 4341 AI 2013a at WPI. All rights are reserved.
*/
package referee;
/*
* @author lzhu
*
*/
public class Board cfw_
int
int
int
int
int
int
int
int
int
int
/
/
width;
height;
board[][];
numOfDiscsInColumn[];
emptyCell=9;
N;
Adding time to Bayes nets:
HMMs and Kalman filters
Recap: Bayesian nets encode conditional
independence relationships
Some oddities of the model
A snap shot of what happened at some point
in time
In a day? A month?
Assumes the past is not important
Cr
Administrivia
Quizzes
Have to be fast since 50 minute class
Takes too long to distribute
Will display on screen instead
Means you must bring paper and pen(cil)
(or hope youre sitting near someone nice)
Getting help on projects
Each project will hav
Intelligent agents
Administrivia
We decided that peer feedback will be
provided immediately after the assignment
Jeff Moffet
jeffmoffet@wpi.edu (syllabus corrected)
Cannot hold office hours this week
Friday, office hours 10:00-11:00 (Elijah)
ADP lab
CS 4341
Introduction to AI
Lecture 1, Introduction
Waitlist
8:00 a.m. section is half full
Priority:
Those who can show me they have a class
conflict at 8:00 a.m.
CS majors and seniors
Office hours
Joseph Beck:
Monday: 9:00 a.m.
Tuesday: 1:00 p.m.
Admissible heuristics
E.g., for the 8-puzzle:
h1(n) = number of misplaced tiles
h2(n) = total Manhattan distance (i.e., no. of squares from desired location
of each tile)
h1(S) = ?
h2(S) = ?
Admissible heuristics
E.g., for the 8-puzzle:
h1(n) = numbe
Informed search
Romania with step costs in km
Adding knowledge: straight-line
distance to Bucharest
SLD: How short would drive be if
didnt need to travel on roads
366
Greedy best-first search
Evaluation function f(n) = h(n) (heuristic)
= estimate of cos
Administrivia
Last chance: going to release assignment #1 soon
Tell me by 5:00 if you have a group
What about using other sources?
Ideas/pseudo code ok
Source code is not
Cite your sources in code, and mention in writeup
Plagiarism / Academic hones
Administrivia
How is assignment #1 going?
Game playing
Games vs. classic search problems
In search, only had to worry about your
actions
In games, opponents moves are often
interspersed with yours
need to consider opponents action
Games typically hav
Administrivia
turnin.cs.wpi.edu is up
Assignment #1 is open for submissions
Due at midnight (test it before then)
Office hours after class until 2:00 (might leave at 1:45 if no one is
there)
Late
20% penalty
Accepted up to Sunday @ 6:00 p.m.
Examp
Administrivia
First quiz on Thursday
Next quiz on Tuesday (holiday on Monday)
Quizzes
Hand something in
Attendance
A point for showing up :-)
Please use normal sized paper
BS/MS
Reminder: let me know if pursuing
Requirements:
A or B
Do addition
Representing independence
Computing conditional probability
What if you want P(Cavity | toothache)
P(C | t) = P(C, t) / P(t)
Note: P(C | t) means we want to know the
probability of each outcome of Cavity
<P(C=true | t), P(C=false | t)>
Joint probability
Reinforcement learning
Reinforcement learning
Long history
Psychology: ~1910 (Skinner)
Computer Science: 1955 (Samuels checkers
player)
Similar to neural nets, interplay between
human and computer learning
Psychology: operant conditioning
RL success
/*
* This code is created for cs 4341 AI 2013a at WPI. All rights are reserved.
*/
package referee;
import java.io.*;
import java.util.*;
/*
* @author lzhu
*
*/
public class Referee cfw_
Board board;
int N=4;
int board_height=6;
int board_width=7;
Timer a
[ part1 mini_max ] | [ part2 alpha_beta ] | [ part3 alpha_beta_limited ]
[ part1 mini_max_clockwise ] | [ part2 alpha_beta_clockwise ] | [ part3
alpha_beta_limited_clockwise ]
red boards are maximizers
blue boards are minimizers
10
o
xoo
x
-10
x o
xoo
x
Homework and Project Goal:
The goal of Project 1 is to help you understand exactly how different search strategies work.
You will implement each of nine net search algorithms. Among the searches are basic searches,
heuristically informed searches, and opt
GAME DESCRIPTION
Tic-Tac-Toe is a two player game (one of them being your computer program). We would deviate
from the classical game which is in 2 dimensions and instead implement a 3-dimensional version
of the game. We would have a 3D board which is a 4
Problem I. Decision Trees (25 points)
Consider the following dataset that specifies the type of contact lenses that is prescribed to a
patient based on the patient's age, astigmatism, and tear production rate. The purpose of this
problem is to use informa
Problem I. Search (20 points)
Suppose that you need to find a path between S and G in the following graph. The number
attached to each edge in the graph represents the COST of traversing the edge. The number
inside each node represents a heuristic under-e
Administrivia
Assignment 2
Final projects
Bayesian inference
Someone has a stiff neck
how likely to be meningitis?
Lets use probability to infer the answer
Observed: stiff neck
Want to know about: meningitis
How do we write probabilistically?
P(me
Administrivia
Past quizzes entered into mywpi
Assignment #2 posted
A lot of points (50%) for non-coding aspects
Optimizing it to do well
Analyzing and writing up results
Finishing code at the last minute will go poorly
Uncertainty
Logical agents
Le
Samuels checkers
Why I adore this paper
Considered two approaches
Large, randomly connected neural networks
Size of network is a problem with 1950s hardware
(were just getting there now)
Vs. special-built approach for just 1 task
Checkers has been sol