Discover related content

Find related publications, people, projects, datasets and more using interactive charts.

View graph of relations

Prof. John Andrew Clark

(Former)

Research interests


Royal Society Wolfson Research Merit Award Holder (2013-2018)

My personal and supervised research work concentrates on aspects of security and software engineering (with a particular interest in applying heuristic computing and other AI to these areas), though I am more eclectic on what I supervise as taught course projects. I was awarded my PhD in 2002 and promoted to a Personal Chair in January 2005. I have been cited (GS) around 4400 times, as far as I can tell.

Summary of My Royal Society Wolfson Research Merit Award Proposal
Computer based systems development is complex, time-consuming, error-prone and costly. Automating much of the process remains a challenging but essential goal. Search Based Software Engineering (SBSE), my primary research area, seeks to couch many problematic design and analysis problems as function optimization tasks. In doing so, it places a raft of research, e.g. from Operations Research and Evolutionary Computation, at the service of software engineering. SBSE offers a vision of how increasing demands on developers can be handled. Our research and the science that underpins it matters to the communities that contribute to it and to the wider software engineering community whose problems we address.

Below I present a short scientific agenda and indicate why it matters. It draws on my SBSE research, starting in classical territory (scheduling), but then seeks to make fundamental contributions outside SBSE’s software engineering and optimization heartlands.

1) Searching for Scheduling. Scheduling mediates access to resources by competing tasks. To date, all scheduling policies have been invented by scheduling theorists. But can search-based approaches discover classic real-time scheduling policies (earliest due date, rate monotonic, static priority pre-emptive scheduling)? Can they discover entirely new and better ones? The discovery of effective policies would, if convincing proofs of correctness could be developed, find ready industrial application, but also spark theoretical work. I will work with Prof A Burns FRAEng, one of the world’s leading scheduling theorists.

2) Discovering Cryptanalysis. Some techniques do not scale with system size. I can find a 16-bit key by brute force but not a 128-bit key. Some techniques do scale. Linear and differential cryptanalysis, for example, are strategies that were applied in the late 1980s/early 1990s to the 56-bit Data Encryption Standard and are now staple attacks against block ciphers. But they can also be launched against “toy” ciphers, e.g. Heys’ tutorial 16-bit block cipher. Toy ciphers may thus serve as the basis from which cryptanalysis techniques could be discovered. I will investigate whether cryptanalysis approaches can be found by search-based techniques (e.g. genetic programming). This will require, for example, the formalization of a “language” in which cryptanalytic techniques can be expressed and a means by expressed techniques can be run, i.e. applied to a set of ciphers and evaluated. This is hard and would constitute a marked increase in the level of sophistication with which AI techniques are used in cryptology. You will have learned long multiplication on quite small numbers, but you could manually multiply two 1000 digit numbers – the technique you learned scales. The same notion applies here. There is a slight irony: the more ambitious the intellectual artefacts sought (analytic strategies) the more evaluation on “toy” case studies becomes relevant. Success will surprise the cryptanalytic community; but this seems an accident waiting to happen. This really would take AI and cryptology into new territory.

3) Finding non-standard computational artefacts. Work in 2004-2006 harnessed evolutionary computation to find useful quantum circuitry and algorithms and won two prizes (GECCO 2004 and 2005). The time is now ripe for further exploitation of automated program discovery in the context of measurement-based quantum computing and in the context of quantum cryptographic protocols. I propose to work with Prof Sam Braunstein (a former WMRA holder) and Dr Stefano Pirandola (a quantum cryptography specialist), and the wider White Rose quantum network. I aim also to continue my recent collaboration on NMR pulse sequence generation with Drs Angelika Sebald and Matthaius Bechmann (described in one of my two highlighted papers). Its results are of genuine interests to NMR specialists, and doing research that matters outside computer science is important to me.

Biography

Royal Society Wolfson Research Merit Award Holder (2013-2018)

I did Mathematics and then an MSc in Applied Statistics at Magdalen College, Oxford. I then joined the security division of the software and systems house Logica in 1987, working as a security and high integrity software evaluation consultant to HMG and then on a security ‘R & D’ thinktank for HMG. I joined York in 1992. I have been research active since around 1997 (with a few outputs before then). My personal and supervised research work concentrates on aspects of security and software engineering (with a particular interest in applying heuristic computing and other AI to these areas), though I am more eclectic on what I supervise as taught course projects. I was awarded my PhD in 2002 and promoted to a Personal Chair in January 2005. I have been cited around 4175 times, as far as I can tell.

Since April 2009 I have been Deputy Head of Department (Responsible for Research). I view my research role largely as helping the people who work with me get on in their careers by producing excellent research. I occasionally think for myself. I am currently engaged in personal research that is exceptionally ambitious but with significant chances of nothing meaningful coming out of it, but then again, if something does....

Most would describe my research work as largely applied, but I have one brain and I have to use it for everything. I don't distinguish too strongly between theory and applied. I am part of the EPSRC's Programme Grant on Dynamic Adaptive Automated Software Engineering (DAASE), led by Mark Harman at UCL.

I am delighted to have received a Royal Society Wolfson Research Merit Award to pursue my work in optimisation based design and analysis of systems.(August 2013-July 2018, announced publicly 30 August 2013).

In the past nine years I have also been co-author of eleven prize-winning research papers: particular thanks to my PhD students, many of whom I taught as undergraduates. Prizes include Evolutionary Computation Human Competitive Awards (Humies) Silver and Bronze medals ($1000 in 2005 and $2000 in 2013) for work on quantum algorithm synthesis (with Massey and Stepney) and NMR pulse sequence generation (with Bechman and Sebald) respectively. I have published or collaborated on proposals with members of every one of our nine research groups.

Many of my PhD students and Research Associates have actually been taught by me. I believe that teaching is an important part of the job and that there is an obligation to communicate the value of science and enegineering  (as well as the fun) to wider audiences. From 2005-2007 I was Chair of Examinations within the Department of Computer Science at York. I was secretary to various Boards of Examiners for around 7 years too. I aim to play as best a part I can in the wider academic community. I have been (taught course) external examiner at eight UK Universities in the past nine years: Open University, Imperial College London, King's College London, Sheffield, and Newcastle have now finished; Brunel, Dublin and University College London are current. I have given various conference keynotes but also open public lectures and lectures to school audiences.