Machine Learning - University of Birmingham

journeycartΤεχνίτη Νοημοσύνη και Ρομποτική

15 Οκτ 2013 (πριν από 3 χρόνια και 11 μήνες)

51 εμφανίσεις

Machine Learning


Bob Durrant

School of Computer Science

University of Birmingham


(Slides: Dr Ata
Kabán
)

Machine Learning: The Module


What is
Learning
?


Decision trees


Instance
-
based learning


Kernel Machines


Probabilistic Models


Bayesian Learning


Learning Theory


Reinforcement Learning


Genetic Algorithms

Lectures & Tutorials


Lectures on Monday at 14.00 in UG40 CS


Tutorials on Thursday at 12.00 in B23 Mech Eng


Exercise sheets given out at lecture


Solutions discussed during tutorials


Handouts are on the module’s web page:
http://www.cs.bham.ac.uk/~durranrj/ML.html

Continuous Assessment


ML: 20% of your final mark


ML
-
EXTENDED: 40% of your final mark


Two types of exercises


Computer based practical work


The exercises are posted on the module’s web page


Deadline: end of term


Paper
-
based exercises (worksheets)


The exercises are on the module’s web page & are handed out in
lectures.


Deadline: before that week’s tutorial session.

Continuous Assessment (cont’d)


Marking:


There will be 12 pieces of assessed work provided during the course.


You must submit at least 6 pieces of work for ML, and at least 10 pieces
of assessed work for MLX.


For MLX, you must submit Practical Assignments 1 and 2 (Assignment 1
counts as 3 pieces of assessed work).


Your assessed work score for ML (resp. MLX) will be the sum of your best
4 (or 8) pieces of submitted work.


Feedback:


You get immediate feedback on Worksheet exercises as we will solve
them in the Thursday tutorial class.


You will also get your marked work returned to you (within 2 weeks).


You can approach me with questions in my office hours (as well as in
tutorials, lectures, breaks).


Office hours


My weekly office hour follows straight after the Monday
lecture, i.e. 15.00


16.00.


You are also welcome to approach me if you see me
around campus.


Location: 134 (First Floor)


What office hours are and aren’t for:


Yes: ask me concrete questions to clarify something
that has not been clear to you from the lecture


Yes: seek advice on your solutions to the given
exercises


Yes: seek advice on further readings on related
material not covered in the lecture


No: ask me to solve the exercises


No: ask me to repeat a lecture

Literature


Machine Learning (Mitchell)


Reinforcement Learning … (
Barto
, Sutton)


Modelling the Web (
Baldi
, Smyth)


Support Vector Machines and Other Kernel
-
Based Learning
Methods (
Cristianini
,
Shawe
-
Taylor)


Artificial Intelligence … (Russell,
Norvig
)


Artificial Intelligence (Rich, Knight)


Artificial Intelligence (Winston)


Elements of Statistical Learning (Hastie,
Tibshirani
, Friedman)


Neural Networks: A Comprehensive Foundation (
Haykin
)

Module Web Page


~durranrj


Syllabus


Handouts


Exercise sheets


Computer
-
based practical exercises


Links to ML resources on the web


Literature

What is Learning?

How can Learning be measured?


Any change in the knowledge of a system that allows it
to perform better on subsequent tasks.


Knowledge. How should knowledge be represented?
Does anybody know how it is represented in the
human brain?


Think for a moment about how knowledge might be
represented in a computer.


If I told you what subjects would come up in the exam,
you might do very well. Would you do so well if I then
set randomly chosen subjects from the syllabus? (This
illustrates the notion called ‘overfitting’
-

something
one should guard against.)

Ways humans learn things


…talking, walking, running…


Learning by mimicking, reading or being told facts


Tutoring


Being informed when one is correct


Experience


Feedback from the environment


Analogy


Comparing certain features of existing knowledge to new
problems


Self
-
reflection


Thinking things in one’s own mind, deduction, discovery


Machine Learning


Interdisciplinary field


Artificial intelligence


Bayesian methods


Computational complexity theory


Control theory


Information theory


Philosophy


Psychology and neurobiology


Statistics




Achievements of ML


Computer programs that can:


Recognize spoken words


Predict recovery rates of pneumonia patients


Detect fraudulent use of credit cards


Drive autonomous vehicles


Play games like backgammon


approaching the
human champion!

What is the Learning problem?

Learning = improving with experience at some task


Improve over task T


With respect to performance measure P


Based on experience E


Example: Learning to play checkers


T: play checkers


P: % of games won in world tournament


E: opportunity to play against self


Example: Learning to recognise faces


T: recognise faces


P: % of correct recognitions


E: opportunity to make guesses and being told
what the truth was


Example: Learning to find clusters in data


T: finding clusters


P: compactness of the groups detected


E: opportunity to see a large set of data

Types of training experience


Direct or indirect


With a teacher or without a teacher


An eternal problem: is the training experience
representative of the performance goal?


It
needs to be.

Forms of Machine Learning


Supervised

learning: uses a series of
examples with direct feedback


Reinforcement

learning: indirect feedback,
after many examples


Unsupervised

learning: no feedback