Spyridon Pougkakiotis
PhD Student at the University of Edinburgh
PhD Student at the University of Edinburgh
My professional interests lie in the field of optimization, and more specifically, large scale optimization. My research is focused on Interior Point Methods (IPMs), Proximal Point Methods (PPMs), Alternating Direction Methods of Multipliers (ADMMs), and their combination. At the same time, I focus on regularization techniques for convex quadratic and semi-definite programming problems as well as on the solution of large scale linear systems of equations using Krylov-subspace methods. All the above combined with a strong background in programming, signal processing and machine learning, form my general research and professional focus.
Apart from my studies, I enjoy spending my time playing and composing jazz and contemporary music, focusing on the electrical or upright bass and the classical piano.
Roughly speaking, my research time is divided as shown below.
As part of my PhD scholarship, I am tutoring in several workshops throughout the academic year. These include workshops from the following courses:
Linear Programming, Modelling and Solution,
I have been a referee for the following Journals:
Numerical Algorithms,
Sep. 2019-present.
SIAM Journal on Optimization,
Apr. 2019-present.
Optimization Methods and Software,
May 2018-present.
Computational Optimization and Applications,
Feb. 2018-present.
Mathematical Programming,
Dec. 2017-present.
I was the main lecturer for the mathematics revision course for students studying an MSc in Statistics, Operational Research or Optimization.
This course included a quick revision of Linear Algebra, Multivariate Analysis and basic Convex Analysis. It consisted of 3 hours of lecturing,
8 hours of workshops and problem solving and 2 written exams.
University - University of Edinburgh
Duration - 1 month [years 2018 and 2019]
Location - Edinburgh, United Kingdom
I am a currently PhD student at the University of Edinburgh. My research interests lie in the field of optimization, and more specifically, developing robust and efficient algorithms for the solution of convex optimization problems. In particular, I am working on the theoretical analysis as well as implementation of Interior Point Methods, Proximal Point Methods, Alternating Direction Methods of Multipliers, and their combinations.
University - University of Edinburgh
First Supervisor - Prof. Jacek Gondzio
Second Supervisor - Dr. John Pearson
Duration - 3.5 years [Sep. 2017 - Jan. 2021]
Location - Edinburgh, United Kingdom
My main focus was in the field of Computational Optimization. Some of the courses attended were:
- Large Scale Optimization for Data Science,
- Modern Optimization Methods for Big Data,
- Non-linear Optimization,
- Stochastic Optimization,
among others.
My dissertation thesis was concerned with the implementation of a Dual Regularized Interior Point Method (IPM), providing either exact or inexact solutions.
University - University of Edinburgh
Grade - Distinction (82.4/100)
Duration - 1 year [Sep. 2016 - Aug. 2017]
Location - Edinburgh, United Kingdom
My studies in the University of Athens were computer science oriented, with my focused areas being signal processing, algorithms and machine learning. (420 ECTS total)
University - National and Kapodistrian University of Athens
Grade - 9.09/10 (Greek Scale), First-class honours (UK Scale)
Duration - 4 years [Spe. 2012 to Aug. 2016]
Location - Athens, Greece
I acknowledge the support I received throughout my studies, by the following sources:
Awarded PhD funding for studying in the Optimization Group of the University of Edinburgh.
Awarded for three consecutive years as additional financial support for my PhD studies in Edinburgh.
Funding for studying an MSc degree at the University of Edinburgh.
A New Preconditioning Approach for an Interior Point-Proximal Method of Multipliers for Linear and Convex Quadratic Programming,
L. Bergamaschi, J. Gondzio, Á. Martínez, J. W. Pearson, S. Pougkakiotis,
arXiv preprint arXiv:1912:10064, 2019.
Fast Solution Methods for Convex Fractional Differential Equation Optimization,
S. Pougkakiotis, J.W. Pearson, S. Leveque, J. Gondzio,
arXiv preprint arXiv:1907.13428, 2019.
An Interior Point-Proximal Method of Multipliers for Convex Quadratic Programming,
S. Pougkakiotis, J. Gondzio,
arXiv preprint arXiv:1904.10369, 2019.
Dynamic Non-Diagonal Regularization in Interior Point Methods for Linear and Convex Quadratic Programming,
S. Pougkakiotis, J. Gondzio,
Journal of Optimization Theory and Applications 181(3), 905-945, 2019.
Efficient KLMS and KRLS algorithms: A Random Fourier Feature Prespective,
P. Bouboulis, S. Pougkakiotis, S. Theodoridis,
2016 IEEE Statistical Signal Processing Workshop (SSP), 1-5.
A workshop focusing on decomposition methods, with a special interest in Alternating Direction Methods of Multipliers (ADMMs).
University - University of Edinburgh
Role - Co-organizer
Duration - 2 days [4-5 June, 2020]
Location - Edinburgh, United Kingdom
Website -
A workshop focusing on numerical linear algebra, with a special interest in partial differential equation optimization.
University - University of Padova
Role - Invited Speaker
Duration - 2 days [17-18 February, 2020]
Location - Padova, Italy
Website -
A major conference taking place every three years, focusing on continuous optimization.
University - Technical University of Berlin
Role - Contributing Speaker
Duration - 6 days [3-8 August, 2019]
Location - Berlin, Germany
Website -
A workshop focusing on numerical linear algebra, with a special interest in matrix-free huge-scale optimization.
University - University of Edinburgh
Role - Invited Speaker
Duration - 2 days [1-2 July, 2019]
Location - Edinburgh, United Kingdom
Website -
A workshop focusing on theoretical and practical developments on continuous optimization.
University - University of Strathclyde
Role - Contributing Speaker
Duration - 2 days [28-29 June, 2019]
Location - Glasgow, United Kingdom
Website -
A seminar focusing on numerical linear algebra.
University - Strathclyde University
Role - Invited Speaker
Duration - 1 day [31 January, 2019]
Location - Strathclyde, United Kingdom
A major conference taking place every three years, focusing on mathematical programming.
University - University of Bordeaux
Role - Contributing Speaker
Duration - 6 days [1-6 July, 2018]
Location - Bordeaux, France
Website -
A conference focusing on numerical linear algebra and partial differential equation optimization.
University - University of Birmingham
Role - Contributing Speaker
Duration - 3 days [27-29 June, 2018]
Location - Birmingham, United Kingdom
Website -
at
sp.pougk at gmail.com
s.pougkakiotis at sms.ed.ac.uk
See social media links