Skip to main content

Novel Application of Advanced Machine Learning Techniques for use in Side Channel Analysis Attacks

CSIT GCHQ studentship

Novel Application of Advanced Machine Learning Techniques for use in Side Channel Analysis Attacks

Principal Supervisor: Máire O'Neill

Project Description:

The Government Communications Headquarters (GCHQ) in Cheltenham has agreed in principle to sponsor a PhD/Doctoral Studentship at Queen's University of Belfast in the area of Novel Application of Advanced Machine Learning Techniques for use in Side Channel Analysis Attacks.

Sensitive information processed by the circuitry in electronic security devices can be leaked via physical characteristics of the device, such as power consumption, electromagnetic (EM) emanation, timing, etc. These techniques are known as Side-Channel Attacks (SCA). To date, a significant amount of research has been carried out into side channel attacks, which uses statistical processing techniques to analyse the information leaked from the device. In this proposed project, we aim to investigate the novel application of advanced machine learning techniques to improve the efficiency and practicality of side-channel analysis.

It is anticipated that this research will illustrate that advanced machine learning techniques can be used to perform successful side channel attacks on protected practical cryptographic implementations, which are currently considered to be resistant to such attacks. The goal is to target current real-world security devices so as to ensure research relevance.

Aims:

The main aims of the proposed research are:

  • To study the state of the art in side channel analysis (SCA) attacks
  • To investigate the application of support vector machines to both software and hardware implementations and, in particular, to look at the impact of parameter selection and how different platforms affect profiling attacks.
  • To examine and compare various machine learning techniques on different platforms in order to determine in what scenarios the different learning techniques have an advantage.
  • To explore unsupervised machine learning techniques as we believe that they have considerable potential in the context of protected implementations.

Academic Requirements:

A minimum 2.1 honours degree or equivalent in Electrical and Electronic Engineering or relevant degree is required.
The studentship is only open to UK nationals and the successful candidate will be required to spend in the region of 2 - 4 weeks per year at GCHQ headquarters in Cheltenham. To be considered for this studentship, candidates must therefore be prepared to undergo GCHQ's security clearance procedures.

General Information

This GCHQ-sponsored PhD studentship provides funding for 3.5 years and commences on 31 September 2013 with a proposed end date of March/April 2017. GCHQ will cover the costs of university fees and will provide an annual stipend to the student corresponding to the National Minimum Stipend (currently £13,590 per annum) plus an additional sum of £7,000 per annum (both tax free). For comparison this is equivalent to approx. £26,555 annual salary.

A further £5k of funding will also be available per annum for travel to conferences, collaborative partners, and GCHQ visits.

Applicants should apply electronically through the Queen's online application portal at: https://dap.qub.ac.uk/portal/

Deadline for submission of applications is 3 May 2013.

(Early submission of applications is recommended due to anticipated interest)

Contact details:

Supervisor Name: Máire O'Neill
Address: ECIT Institute,
Queens Road,
Belfast,
BT3 9DT
Email: m.oneill@ecit.qub.ac.uk
Tel: +44 28(0) 9097 1785
Web: http://www.csit.qub.ac.uk

Deadline for Submission of Applications: 3rd May 2013

For further information on Research Area click on link below:
http://www.csit.qub.ac.uk/