School of Engineering and Informatics (for staff and students)

Junior Research Associates

Applications for the 2022 JRA programme are now open.

The deadline to submit a JRA application is 12.00 noon on 28th March 2022.

Junior Research Associates 2022

The University of Sussex Junior Research Associate (JRA) scheme is a pioneering project which aims to develop future research leaders. It encourages talented and ambitious undergraduates to consider a career in research following graduation. Each JRA undertakes an intensive eight-week, full-time research project. This usually takes place in their second summer of study. 

The JRA scheme seeks to benefit students from any background and group. We particularly want the JRA to be accessible to students who might not otherwise be exposed to research experience or to consider a career in academic research. The Doctoral School therefore offers a number of JRA bursaries worth £1,800 for living costs plus £200 for research expenses to help our students study full time over the summer. 

Students can either apply with one of the projects below or suggest an original proposal (as long as a member of faculty is willing to supervise).

For further information about the scheme and how to apply visit the JRA webpages.

Applicants must be sponsored by a member of faculty so please contact the project supervisor to dicuss the project and application. Visit our website for a full list of our Informatics faculty

See below for current projects within the Department of Informatics. Click on the project title to reveal a description:

Distributed graph processing in the Kuramoto model of synchronisation

Project Supervisor: Prof Luc Berthouze

The Kuramoto model [1] is often used to study synchronisation in the brain. In its simplest form, the model assumes a fully connected network, but it can be adapted to operate on more realistic networks. We are interested in understanding how network structure impacts synchronisation. This involves running a large number of simulations using different network structures. For large networks, this is computationally intensive. The goal of this JRA project is to explore the use of a distributed graph processing platform [2] such as Giraph to accelerate this process. Strong experience coding in Java and interest in distributed processing is essential.

[1] See: http://tutorials.siam.org/dsweb/cotutorial/for an easy introduction. For more detail, see Acebrón et al. (2005). The Kuramoto model: A simple paradigm for synchronization phenomena. Reviews of modern physics. 77(1):137. Available at: https://www.researchgate.net/publication/46776356_The_Kuramoto_model_A_simple_paradigm_for_synchronization_phenomena.

[2] Malewicz et al. (2010). Pregel: a system for large-scale graph processing. In Proceedings of the 2010 ACM SIGMOD International Conference on Management of data (pp. 135-146). Available at: https://www.cs.cmu.edu/~pavlo/courses/fall2013/static/papers/p135-malewicz.pdf

NLP meets temporal logics

Project Supervisor: Dr Hsi-Ming Ho

Recently there have been some attempts to apply modern NLP / machine learning techniques to traditional logical/mathematical reasoning tasks, e.g., simple arithmetic or Boolean formula satisfiabilty. A recent work, which applies the Transformer architecture to the problem of finding solutions to temporal logic formulas, can be found at the link below:

https://openreview.net/forum?id=dOcQK-f4byz

The goal of theproject is to advance the state of the art by applying similar techniques to more expressive temporal logics (e.g., with timing constraints). It is likely that the result will degrade as the formulas will be more complicated, thus it is also expected thatone needs some novel ideas to improve the performance.

Document layout analysis for exam scripts

Project Supervisor: Dr Hsi-Ming Ho

Document image analysis (DIA) plays a key role in modern social sciences and humanities research. While there have been many success stories in various individual projects based on recent advances in using deep learning for DIA, many of these required specialised training or post-processing steps to achieve satisfactory performances. The goal is to make use of a recently proposed uniform framework for DIA [1] to develop a semi-automated workflow for exam marking -based on the assumption that scanned PDF exam scripts have rather restricted (and known) possible layouts.

[1] Shen et al. LayoutParser: A Unified Toolkit for Deep Learning Based Document Image Analysis. https://arxiv.org/pdf/2103.15348.pdf

Evaluating cryptographic algorithms for secure content dissemination schemein the Internet of Vehicles

Project Supervisor: Dr Naercio Magaia

Security is a key challenge on the Internet of Vehicles (IoV) since, nowadays, many security attacks exist that can negatively impact its reliability [1]. Some examples of common threats to vehicles in IoV include denial of service, jamming, the man in the middle, Sybil, eavesdropping, and broadcast and message tampering.

The volume of data required by connected vehicles in IoV will continue to rise, and some of the appealing vehicle content dissemination applications are of upmost importance for the correct functioning of vehicles [2].

This project aims to evaluate novel cryptographic algorithms used for secure content dissemination to avoid attackers from tampering with critical data that would endanger vehicles’ passengers on the Internet of Vehicles.

[1] C. Bernardini, M. R. Asghar, and B. Crispo, “Security and privacy in vehicular communications: Challenges and opportunities,” Vehicular Communications, vol. 10. pp. 13–28, 2017.

[2] N. Magaia, Z. Sheng, P. R. Pereira, and M. Correia, “REPSYS: A robust and distributed incentive scheme for in-network caching and dissemination in Vehicular Delay-Tolerant Networks,” IEEE Wirel. Commun. Mag., pp. 1–16, 2018.

Model Checking Properties of Programmable Networks

Main supervisor: Dr Bernhard Reus

Software Defined Networking (SDN) is a new paradigm for operating and managing computer networks. SDN allows anyone to write SDN applications but this could potentially result in poorly written and very buggy software. It is therefore crucial to develop techniques for verifying the correctness of software-defined network functionality like absence of black holes or loops. This project investigates whether network properties can be verified with the model checking tool ESBMC. More precisely, using simple examples, investigate: 1) How to best model a SDN network as a concurrent C program? 2) Can ESBMC handle those C programs to verify important safety properties?

[1] L. Cordeiro and B. Fischer. Verifying Multi-Threaded Software using SMT-based Context-Bounded Model Checking. ICSE, pp. 331–340, 2011. https://ieeexplore.ieee.org/document/6032472

[2] Klimis, Vassilis, Parisis, George and Reus, Bernhard: Towards model checking real-world software-defined networks. In: Computer Aided Verification. CAV 2020. Lecture Notes in Computer Science. 12225 126-148. Springer Verlag. http://sro.sussex.ac.uk/id/eprint/90723/5/Klimis2020_Chapter_TowardsModelCheckingReal-World.pdf

[3] ESBMC online manual: https://ssvlab.github.io/esbmc/documentation.html

Software to Support the Understanding of HLA-type-associated Disease Risk

Project Supervisor: Dr. Bernhard Reus

Genetically encoded HLA-molecules present ‘self’ and 'foreign' peptides to T-lymphocytes, an important immune cell subset. This may result in protective but also self-destructive (auto)immunity. Therefore, the concrete HLA allele (i.e. gene variation) of a patient may impact their susceptibility to infection and severity of symptoms.
You will support an interdisciplinary research team of computer scientists and immunologists developing a software tool to find novel, statistically relevant associations between HLA alleles and viral diseases like CMV and COVID-19.You will design or write prototype software components to filter data, visualise data, or compute statistical models, depending on your experience. 
[1] S.C.L Gough and M.J Simmonds. The HLA Region and Autoimmune Disease: Associations and Mechanisms of Action. Curr Genomics. 2007 Nov; 8(7): 453–465 (2007). https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2647156/
[2] Yanhui Fan and You-Qiang Song. PyHLA: tests for the association between HLA alleles and diseases. BMC Bioinformatics volume 18, 90 (2017). https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-017-1496-0
[3] Douillard Venceslas et al.: Approaching Genetics Through the MHC Lens: Tools and Methods for HLA Research, Frontiers in Genetics 12, 2021.https://www.frontiersin.org/articles/10.3389/fgene.2021.774916/full[4] Douillard Venceslas et al.: Current HLA Investigations on SARS-CoV-2 and Perspectives, Frontiers in Genetics 12, 2021. https://www.frontiersin.org/articles/10.3389/fgene.2021.774922/full
Learning an interpretable model for editing images

Project Supervisor: Dr Ivor Simpson

Image to image translation models, such as CycleGAN[1], provide a framework for transforming an image from one domain (e.g. horses) to another (e.g. zebras). These methods have enjoyed substantial success in being applied to various domains, however in general the translation functions are overly flexible and lack inspectability, which can result in undesirable changes.

This project would use an iterative neural renderer [2][3], which creates/edits an image through the application of sequential virtual brush strokes as the translation function, this could be extended to model shape changes through geometric transformations [4]. These interpretable editing techniques provide reasonable limits to the complexity of changes and provide a mechanism for human inspection.

[1] Zhu, Jun-Yan, et al. "Unpaired image-to-image translation using cycle-consistent adversarial networks." Proceedings of the IEEE international conference on computer vision. 2017.

[2] Huang et al., Learning to Paint With Model-based Deep Reinforcement Learning, Proceedings of the IEEE international conference on computer vision. 2019.

[3] Ha, David, and Douglas Eck. "A neural representation of sketch drawings." arXiv preprint arXiv:1704.03477 (2017).

[4] Dorta, Garoe, et al. "The GAN that warped: Semantic attribute editing with unpaired data." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2020

Modelling heterogeneity and uncertainty in neurodegenerative disease progression

Project Supervisor: Dr Ivor Simpson

Neurodegenerative diseases, such as Alzhiemer's, are known to have heterogeneous presentation across populations. This can complicate differential diagnosis between various forms of dementia, and predicting how a patient's state may degrade over time.

This project will use a public dataset[1,2] containing neurological biomarkers to build machine learning models, such as [3,4,5], that predict disease state/progression give a set of measurements. You will investigate methods to quantify the uncertainty of predictions and better understand population variability.

[1]https://tadpole.grand-challenge.org/

[2] http://adni.loni.usc.edu/

[3]https://github.com/ucl-pond/pySuStaIn

[4] Kingma, Diederik P., and Max Welling. "Auto-encoding variational bayes." ICLR 2013[5] Vaswani, Ashish, et al. "Attention is all you need." Advances in neural information processing systems. 2017.

Developing Efficient Network Protocols for Low-Earth Orbit Satellite Networks

Project Supervisor: Dr George Parisis

Thousands of Low Earth Orbit (LEO) satellites currently orbit the Earth,which are major components of vast constellations with the goal of 100%geographic coverage. Various private companies (e.g., SpaceX, OneWeb)currently lead the race with billion-dollar investments in place [1]. Theseconstellations exhibit a unique combination of characteristics [2, 3, 4]; (1)aggregate bandwidth in the order of hundreds of Tbps, which is compara-ble to today’s aggregate fibre capacity [2]; (2) multiple paths in a denseand constantly changing network topology; (3) sub-10ms round-trip timebetween the Earth and the first-hop satellite; (4) low but varying end-to-end latency that can be smaller than what the best theoretical fibre pathcan support. With these characteristics being collectively unprecedented,a major rethink of routing and data transport mechanisms are needed toensure efficient usage of network resources.

This JRA project is part of a broader research effort in devising nextgeneration data transport for LEO satellite constellations. The project willinvolve using a LEO satellite constellation simulation model [5], writtenin C++ using the OMNeT++ framework [6], developed internally by thePackets Lab.

References

[1] Tobias Klenze, Giacomo Giuliari, Christos Pappas, Adrian Perrig, andDavid Basin. Networking in Heaven as on Earth. InProceedings of the17th ACM Workshop on Hot Topics in Networks, pages 22–28, RedmondWA USA, November 2018. ACM.

2] Debopam Bhattacherjee, Waqar Aqeel, Ilker Nadi Bozkurt, AnthonyAguirre, Balakrishnan Chandrasekaran, P. Brighten Godfrey, GregoryLaughlin, Bruce Maggs, and Ankit Singla. Gearing up for the 21st cen-tury space race. InProceedings of the 17th ACM Workshop on Hot Topicsin Networks, pages 113–119, Redmond WA USA, November 2018. ACM.

[3] Mark Handley. Delay is Not an Option: Low Latency Routing in Space.InProceedings of the 17th ACM Workshop on Hot Topics in Networks,pages 85–91, Redmond WA USA, November 2018. ACM.

[4] Mark Handley. Using ground relays for low-latency wide-area routing inmegaconstellations. InProceedings of the 18th ACM Workshop on HotTopics in Networks, HotNets ’19, pages 125–132, New York, NY, USA,2019. Association for Computing Machinery.

[5] Aiden Valentine and George Parisis. Developing and experimenting withLEO satellite constellations in OMNeT++. InProceedings of the 8thOMNeT++ Community Summit 2021, September 2021.

[6] OMNeT++. https://omnetpp.org/.

Conversational AI to support Python Debugging

Project Supervisor: Dr Kate Howland

This project is a collaboration with My Code Kit, an Innovate UK backed SME who are developing Ocobox [1] - an AI learning assistant and Python IDE which encourages novice programmers to experiment and fail with confidence. This project involves prototyping conversational interactions for Ocobox including the Python responses and identifying initial sentiments towards them. It forms part of a wider research programme around the use of natural language interactions to support novice programmers [2, 3].

JRA project work will involve:

  • Collecting regular Python error messages and common feelings/sentiments towards them. (This could potentially be done by scraping public conversations on Twitter and other online sources such as Stack Overflow/ Reddit).
  • Modelling conversational responses to Python Error Messages (NLP)
  • Prototyping conversational AI
  • Collaborating with software engineers to serve machine learning models in production

[1] www.mycodekit.com

[2] Good, Judith, and Kate Howland. Programming language, natural language? Supporting the diverse computational activities of novice programmers. Journal of Visual Languages & Computing 39 (2017): 78-92. https://doi.org/10.1016/j.jvlc.2016.10.008

 

[3] Howland, K., & Good, J. (2015). Learning to communicate computationally with Flip: A bi-modal programming language for game creation. Computers & Education, 80, 224-240. https://doi.org/10.1016/j.compedu.2014.08.014

School of Engineering and Informatics (for staff and students)

School Office:
School of Engineering and Informatics, University of Sussex, Chichester 1 Room 002, Falmer, Brighton, BN1 9QJ
ei@sussex.ac.uk
T 01273 (67) 8195

School Office opening hours: School Office open Monday – Friday 09:00-15:00, phone lines open Monday-Friday 09:00-17:00
School Office location [PDF 1.74MB]