Graphical Models
Course Material
Lectures
Please note that while lecture notes will be provided, the lecture itself
will sometimes digress to web-based material and Java applets for
demonstrations. These will usually be listed on this page, but
not on the PDF slides.
Lecture 1: Introduction: Graphs and Probability
Review of some famous graphical models including circuits, fault
trees, and ising lattices. Review of graphs including definitions for
directed and undirected graphs, relations such as parents and
ancestors, and connectivity notions. Review of probability as
frequency or belief in both discrete and continuous contexts.
Review of modelling as a key part of intelligent systems design.
Examples of using graphs to model dependence, causality and relevance.
[ Lecture 1 in PDF,
in Postscript, 2-up, gzipped ]
Preparatory reading: Please review the material
under graphs from Graph Theory and the Ising Model, and course notes of
Three Concepts: Probability
(PDF).
Lecture 2: Independence Modelling
Functional independence in optimization as a method for task
simplification. Graph partitioning as a means of problem
decomposition. Probabilistic independence as functional independence
in probability space. Undirected graphs, relevance, and independence
defined as set separation. Directed graphs, dependence, and
independence defined on the underlying undirected graph. Soundness
and completeness between graphical independence and functional
independence forms. A catalogue of graphical forms for time-series,
spatial modelling, diagnosis and fault detection and their
independence properties. [Lecture 2 in PDF
,
in Postscript, 2-up, gzipped ]
Preparatory reading: Please review
Shachter's manuscript "Examples without Numbers" (handed out in Lecture 1) and Scheines'
tutorial on d-separation. You don't have to understand d-separation,
because its difficult and we'll use a Lauritzen-style definition instead,
but have a look at it.
Part I of Bishop's tutorial (see below) is good material for this lecture.
Lecture 3: Exact Inference
Given a graphical model on discrete variables of low dimension, and a
probability formula, evaluate the probability, or find the maximum
likelihood assignment for the free variables. We will consider this
in a general sense.
We will develop an exact algorithm for the computation as nested
summation on a clustered graph using variable elimination as the basic step.
This turns out to create a clique tree, the general independence structure.
Thus we will consider finding good clique trees, and their complexity.
We will not cover implementation details for the algorithm used to
speed it up in practice.
[Notes in PDF (updated since lecture!!),
in Postscript, 2-up, gzipped]
- Variable elimination demo (not currently available).
Preparatory reading: Please review R. McEliece and S. M. Aji, 2000. The Generalized Distributive Law,
IEEE Trans. Inform. Theory, vol. 46, no. 2 (March 2000), pp. 32
5--343. Read sections I, II and VI only.
Lecture 4: Introduction to Modelling and Learning
We begin by viewing B-Course applied to the "Visit to Asia" network.
Learning will look at simple cases: learning proportions, and learning a
regression curve. These examples are covered in detail because they
form the basis of many algorithms.
The proportions example will be done using the
excellent Beta Coin Experiment applet from the Prob./Stats Object Library.
They introduce the notion of sufficient statistics and the trade-off between
complexity and predictive accuracy, and they illustrate the problems with priors.
Discussion of modelling looks at the interaction between design
decisions and the quality of the developed/proposed solution.
[Notes in PDF (updated since lecture!!),
in Postscript, 2-up, gzipped]
Preparatory reading: Please review Henry Tirri's notes on
proportions.
Heckerman's tutorial is a reference for this section.
Lecture 5: Learning with Graphs
Learning of graphs requires one to learn both the
graphical structure and the probability parameters at each node.
Probabilistic reasoning can handle both, given suitable assumptions.
Learning on graphs will look at methods in the
B-Course system.
We will look at the maximum likelihood approach and some priors
used for the Bayesian approach. The search space will
be briefly discussed. The key techniques used here will
be considered: parameter independence, the exponential family,
the notion of "evidence" and its use, and multiple models.
[Notes in PDF (updated since lecture!!),
in Postscript, 2-up, gzipped
Preparatory reading: Please look at the B-Course Library.
You could optionally review initial part of Friedman and Goldszmidt's tutorial,
not beyond first 50 slides (first 20 is review).
Have a practice on B-Course with some example data generated
using Teemu's generator.
Lecture 6: Selected Applications
Applications will be presented by Tomi Silander on learning with B-Course,
on new PCA models in information retrieval by Wray Buntine, and on
location sensing by Petri Myllymäki.
Preparatory reading: None for this week.
Literature, further reading
There is no need to purchase any books. We will
supply copies of any necessary material prior to lectures. Material
available online will sometimes be in Adobe's PDF or Postscript format
so you should be able to view these formats. Moreover, some of the
demonstrations require Java capable browsers. Other than the use of
these long-time standard formats, we make efforts to avoid online
material that requires special browsers (e.g., thus material converted
from Powerpoint is avoided). Note when we say something is a
reference book, we mean you should not use it to learn the
material, but only to check details or find more advanced material.
Graphs
-
Graph Theory from Math World
including
graph,
directed acyclic graph,
forest,
tree,
degree,
outdegree and indegree,
clique,
induced subgraph, and
connected
component.
Some additional material we will touch on and you probably wont know
is
cut point,
mincut,
minimum spanning tree,
random graph, and phase transition.
What is missing strangely is the notion of width and
parent or ancestor relations.
-
Ising Model (1)
Ising Model (2)
is an undirected graph in a grid where all points are
connected to their horizontal and vertical neighbors. It works like a
probabilistic version of the game of life.
-
“Why is Automatic Test Pattern Generation (ATPG) easy?”
(PDF, 627kB) subject is
nothing to do with this course, but the paper illustrates the notions
of tree-width using a heuristic proxy called "cut-width" (a safe upper
bound) to tree-width we define later. The hardness of problems in
practice is shown as a function of this in a real-world context. Skim
read this; you aren't supposed to understand the details.
- “A tourist guide through treewidth”
(PDF, 1311kB), by Bodlaender. Accessible but
much more detail than we need. Yet shows you the breadth and use of
the computational/optimization side of graphical modelling.
Probabilistic graphs get a mention under the old tag of “expert
systems”.
Probability
Graphical Models
-
“Probabilistic Graphical Models Part One: Graphs and Markov Properties
”
(PDF, 890kB),
and “Probabilistic Graphical Models Part Two: Inference and Learning
”
(PDF, 1394kB), by C.M. Bishop.
Good tutorial overlapping quite a lot of the course.
-
Heckerman's
tutorials,
especially “A Tutorial on Learning with Bayesian
Networks” and “Technical overview for statistician (includes
undirected graphs) - slides from Valencia 2002 tutorial”.
Used in the learning lectures.
-
Friedman's
AAAI-98 tutorial with Goldszmidt,
covers different material to his NIPS-2001 tutorial,
though the NIPS-2001 tutorial is more recent.
- A Brief Introduction to Graphical Models and Bayesian Networks, Kevin Muphy's
tutorial from 1998.
-
B-Course developed in the department.
Used in the learning lectures.
- Learning in Graphical Models edited by M.I. Jordan, 1999
(available in Computer Science Department library). An advanced
reference book and a good source of term paper material. Individual
chapters also
available on-line.
- M.I. Jordan, “Graphical Models”
(PDF, 1239kB), Statistical Science, 2003.
- Graphical Models by S.L. Lauritzen, 1996 (available in
Computer Science Department library). An advanced reference book.
Lauritzen has been instrumental in a lot of early developments.
- R. McEliece and S. M. Aji, 2000. The Generalized Distributive Law, IEEE Trans. Inform. Theory, vol. 46, no. 2 (March 2000), pp. 325--343. Overview and reference for exact computation.
- Graphical Models in Applied Multivariate Statistics by
Whitaker, 1990 (available in Valtiotieteellisen
tiedekunnan kirjasto). A good reference book from the statistical
side, intended as an advanced text for graduates.
- “A catalogue of graphs”
(PDF, 1318kB),
by Studeny.
Ignore the first 8 pages of theory, the rest catalogues
different types of directed and undirected graphs over 2-5 variables.
In general, Studeny is an excellent source of theoretical
work on independence models, probability and graphs.
- Graphical Models Discussion Papers edited by Joel Chenu, another reference, collection
of research-oriented tutorials.
-
Cecil Huang and Adnan Darwiche.
Inference in belief networks: A procedural guide.
(PDF, 451kB)
International Journal of Approximate Reasoning, 15(3):225-263,
October, 1996.
Toolboxes/Software