Postgraduate Course: Numerical Methods for Data (MATH11240)
Course Outline
School  School of Mathematics 
College  College of Science and Engineering 
Credit level (Normal year taken)  SCQF Level 11 (Postgraduate) 
Availability  Available to all students 
SCQF Credits  10 
ECTS Credits  5 
Summary  A datafitting/approximation theory/numerical analysis view of modern computational techniques in data science and machine learning. Focuses on algorithms and the maths behind those algorithms. Complements other courses in the Schools of Mathematics and Informatics that focus on statistical viewpoints and/or implementation and testing. Draws on ideas from applied linear algebra and matrix computation. 
Course description 
Recap: Matrix Computation
Topic 1: Networks
Topic 2: Inverse Problems
Topic 3: Deep Learning with Artificial Neural Networks
Topic 4: Adversarial Attacks

Entry Requirements (not applicable to Visiting Students)
Prerequisites 
Students MUST have passed:
Numerical Linear Algebra (MATH10098)

Corequisites  
Prohibited Combinations  
Other requirements  None 
Information for Visiting Students
Prerequisites  None 
High Demand Course? 
Yes 
Course Delivery Information

Academic year 2023/24, Available to all students (SV1)

Quota: None 
Course Start 
Semester 2 
Timetable 
Timetable 
Learning and Teaching activities (Further Info) 
Total Hours:
100
(
Lecture Hours 22,
Seminar/Tutorial Hours 5,
Supervised Practical/Workshop/Studio Hours 6,
Summative Assessment Hours 2,
Programme Level Learning and Teaching Hours 2,
Directed Learning and Independent Learning Hours
63 )

Assessment (Further Info) 
Written Exam
50 %,
Coursework
50 %,
Practical Exam
0 %

Additional Information (Assessment) 
Coursework: 50%
Exam: 50% 
Feedback 
Not entered 
Exam Information 
Exam Diet 
Paper Name 
Hours & Minutes 

Main Exam Diet S2 (April/May)  Numerical Methods for Data (MATH11240)  2:00  
Learning Outcomes
On completion of this course, the student will be able to:
 Use spectral (eigenvector/eigenvalue) information in data analysis.
 Demonstrate understanding of the links between matrix computation (applied linear algebra) and network science (applied graph theory).
 Demonstrate understanding of mathematical concepts needed to design and train a deep learning network.
 Quantify success in a classification task.
 Apply numerical algorithms to problems in data science, including clustering problems, inverse problems and classification problems.

Reading List
Illustrative Resources:
Matrix Computation:
Linear Algebra and Learning from Data, G. Strang, WellesleyCambridge Press, 2019.
Networks:
A First Course in Network Theory, E. Estrada and P. A. Knight, Oxford, 2015.
Inverse Problems:
Discrete Inverse Problems: Insight and Algorithms, P. C. Hansen, SIAM, 2010.
Deep Learning:
Deep learning: an introduction for applied mathematicians,C. F. Higham and D. J. Higham, SIAM Review, 860¿891, 2019.
Deep Learning, I. Goodfellow, Y. Bengio and A. Courville, MIT Press, 2016.
Adversarial Attacks:
Intriguing properties of neural networks, C. Szegedy et al., Int Conf Learning Representations, 2014. 
Additional Information
Graduate Attributes and Skills 
Not entered 
Keywords  numerical,methods,data 
Contacts
Course organiser  Dr Kostas Zygalakis
Tel: (0131 6)50 5975
Email: K.Zygalakis@ed.ac.uk 
Course secretary  Miss Gemma Aitchison
Tel: (0131 6)50 9268
Email: Gemma.Aitchison@ed.ac.uk 

