We have now dispensed with the necessary background material for AI problem solving techniques,
and we can move on to looking at particular types of problems which have been addressed using AI
techniques. The first type of problem we'll look
6.1 There's Reasoning, and then There's
As humans, we have always prided ourselves on our ability to think things through: to reason things out
and come to the only conclusion possible in a Sherlock Holmes kind of way. But what
Search in Problem Solving
If Artificial Intelligence can inform the other sciences about anything, it is about
problem solving and, in particular, how to search for solutions to problems. Much of
AI research can be explained in terms of specifying a probl
To recap, we now have some characterizations of AI, so that when an AI problem arises, you will be able
to put it into context, find the correct techniques and apply them. We have introduced the agents language
so that we can talk
Two Player Games
One player loses, one player wins
One player wins what the other one loses
See game theory for the mathematics
rather than cooperative
an agent to play a game
Artificial Intelligence Agents
This topic is intended to introduce the fundamental concepts and terms that will
be used throughout this AI class.
Autonomous Rational Agents
Search in Problem Solving
Problem Solving Agents
Wants environment to be in particular state
a number of possible actions
An action changes environment
to satisfy some goal
sequence of actions reaches the goal?
Introduction to AI
Italy, 13th April 2008. A journalist stops three friends (Mr. Rossi, Mr. Bianchi and
Mr. Verdi) who just left the voting office. The journalist interviews them
separately about their votes in the elections. Mr. Rossi declares: If Bianch
Introduction to Artificial
Characteristics of AI Systems
can be broadly listed as follows;
Ability to work with Incomplete data
Ability to work with Conflicting Data
o A C+ project is spread over several files.
o Only one file contains a main() function.
o Before using a function/class
Amount and types of parameters.
o have to be known, but not implemented!
o Definitions can be in
5 Wrapping up the pipeline
In terms of mapping, - 1 is an item that is close to the camera, and 1 is far away from the camera.
The perspective frustum is mapped on a cube, as all values are between 1 and 1.
o This test is known as a clipping test.
o We cu
7 Ray Tracing
o A ray in space is defined by
o Any point along the ray is r(t) = o + t * d.
T > 0: Point in front of the origin.
T < 0: Point behind the origin.
o The most important task in ray tracing is to i
4 Let there be light / light stuff / shading
Images are not yet interesting/exciting, due to an obvious lack of light and shading.
Color is perceived depending on the spectral power distribution of the light, which, when mixed
with an objects reflectance
2 Graphics pipeline
Projection Transform coordinates to the screen.
The depth test is a method to render depth items in rasterization by only displaying the closest
triangles that are recorded.
Virtual camera model
o Graphical movement is not so much the
1 Introduction, painting by numbers
The midterm after the first five lectures does not cover the practical work, only the theoretical
o 50% the midterm exam.
o 35% the project.
o 10% certain practical work, these will not be announced.
8 Ray Tracing, part 2
Reflection and refraction tradeoff
o Schlicks approximation = F(o1) = F0 + (1 F0)(1- cosO1) 5
o F0 is the constant, the reflection for the upright light.
Ray Tracing is slow
o Runs in O(nm), where n = objects and m = rays.
o Even wit
Hasenfratz: Shadow is the region of space for which at least one point of the light source is
Umbra region = An area where a light source is not visible / pitch black.
Penumbra region = An area where the light source can be partially s