AC51025 Probabilistic Inference and Machine Learning 12-13x

stemswedishΤεχνίτη Νοημοσύνη και Ρομποτική

15 Οκτ 2013 (πριν από 4 χρόνια και 29 μέρες)

105 εμφανίσεις

University of Dundee

Module Specification


Code

AC51025



Title

Probabilistic Inference and Machine Learning


College

CASE


School

Computing


Credit rating
: l
evel
, no. of credits

SHE M (SCQF 11), 20 SCQF credits = 10

ECTS credits


Aims

To provide students

with knowledge and

understanding of
methods for probabilistic modelling,
learning and inference. To equip students with a range of
principled
methods for solving

complex

real
-
world
problems
in the presence of uncertainty.


Intended learning outcomes

KNOWLEDGE about:



Commonly used probabilistic models



Methods for performing inference

and machine learning



Algorithms for density estimation, regression and classification



Learning in

Bayesian networks and neural networks

UNDERSTANDING of:



How to perform
consistent reasoning in the presence of uncertainty



How to formulate models and perform inference in a principled way to solve complex
problems

SKILLS in:



Implemen
ting and experimenting with probabilistic models



Indicative content



Introduction to prob
ability
calculus



Probabilistic inference, enumeration



Maximum likelihood, EM, mixture models, clustering, useful distributions



Model comparison and Occam’s razor



Marginalisation, exact inference, approximate in
ference (Laplace’s method, MCMC
)



Latent varia
bles, hidden Markov models, Bayesian networks



Neural networks


Modes of delivery & student participation

The module mixes conventional face
-
to
-
face, lecture
-
based delivery (accompanied by slides) and
tutorials with hands
-
on practical work using appropriate software.


Teaching
,

learning

and assessment

This
module includes approximately 25 hours of lectures an
d 25

hours of other meetings for lab
-
based exercises or tutorials.


Summative a
ssessment
:
Coursework (%)

Exam
ination

(%
,
No. & duration of exam
s.
)

Coursework (20%) and Examination (8
0%, a 2

hour examination)


When taught

S1


Pre
-
requisites or

entry requirements

N/A

Co requisites

None

Antirequisite
s

None

Further information

None