Skip to main content
School of Electronic Engineering and Computer Science

CSC PhD Studentships in Electronic Engineering and Computer Science

About the Studentships 

The school of Electronic Engineering and Computer Science of the Queen Mary University of London is inviting applications for several PhD Studentships in specific areas in Electronic Engineering and Computer Science co-funded by the China Scholarship Council (CSC). CSC is offering a monthly stipend to cover living expenses and QMUL is waiving fees and hosting the student. These scholarships are available only for Chinese candidates. For details on the available projects, please see below.

About the School of Electronic Engineering and Computer Science at Queen Mary 

The PhD Studentship will be based in the School of Electronic Engineering and Computer Science (EECS) at Queen Mary University of London. As a multidisciplinary School, we are well known for our pioneering research and pride ourselves on our world-class projects. We are 8th in the UK for computer science research (REF 2021) and 7th in the UK for engineering research (REF 2021). The School is a dynamic community of approximately 350 PhD students and 80 research assistants working on research centred around a number of research groups in several areas, including Antennas and Electromagnetics, Computing and Data Science, Communication Systems, Computer Vision, Cognitive Science, Digital Music, Games and AI, Multimedia and Vision, Networks, Risk and Information Management, Robotics and Theory. 

For further information about research in the school of Electronic Engineering and Computer Science, please visit: http://eecs.qmul.ac.uk/research/. 

Who can apply 

Queen Mary is on the lookout for the best and brightest students. A typical successful candidate:  

  • Should hold, or is expected to obtain an MSc in the Electronic Engineering, Computer Science, or a closely related discipline 
  • Having obtained distinction or first-class level degree is highly desirable 

Eligibility criteria and details of the scheme 

https://www.qmul.ac.uk/scholarships/items/china-scholarship-council-scholarships.html 

How to apply 

Queen Mary is interested in developing the next generation of outstanding researchers and decided to invest in specific research areas. For further information about potential PhD projects and supervisors please see below. 

Applicants should work with their prospective supervisor and submit their application following the instructions at https://www.qmul.ac.uk/eecs/phd/how-to-apply/

Applicants are also advised to check CSC guidelines and requirements, and to prepare all the documentation well in advance. For example, applicants who are currently employed in China may need a letter of support from their employer, which may take time to obtain.

The application should include the following: 

  • CV (max 2 pages)  
  • Cover letter (max 4,500 characters) stating clearly in the first page whether you are eligible for a scholarship.  
  • Research proposal (max 500 words) 
  • 2 References  
  • Certificate of English Language (for students whose first language is not English)  
  • Other Certificates  

Application Deadline 

The deadline for applications is the 27th January 2027, 4pm. 

For general enquiries contact Mrs. Melissa Yeo m.yeo@qmul.ac.uk (administrative enquiries) or Dr Arkaitz Zubiaga a.zubiaga@qmul.ac.uk (academic enquiries) with the subject “EECS-CSC 2027 PhD scholarships enquiry”. 

Specific projects are listed below.

Supervisors: Akram Alomainy and SaeJune Park

This PhD would design and evaluate capsule antennas and wearable RF receivers to improve capsule tracking and data communication inside the body. It would study antenna performance and signal propagation, then use machine learning to interpret RF measurements and estimate the capsule’s location in real time. The work would build on CE-Track’s wearable tracking approach and QMUL’s wireless capsule endoscopy antenna research.

Supervisor: SaeJune Park

Terahertz (THz) spectroscopy has emerged as a powerful non-destructive tool for label-free biomolecular analysis, offering distinct spectral identification of complex biological systems. However, examining trace amounts of biomolecules in physiological or aqueous environments remains challenging due to strong water absorption and weak THz light–matter interactions at sub-wavelength scales.

This PhD project aims to develop novel microfluidic-integrated 3D THz metamaterial platforms to achieve ultra-sensitive, real-time biosensing. By exploiting three-dimensional electromagnetic field confinement, the proposed architectures will significantly enhance the interaction volume and local field intensity within microfluidic channels. The successful candidate will investigate the numerical design, microfabrication, and THz time-domain spectroscopy characterisation of integrated 3D microfluidic–metamaterial sensors. This interdisciplinary research will provide fundamental insights into high-frequency dielectric dynamics and establish a robust, low-volume sensing route for early disease diagnosis and environmental biomonitoring.

Supervisor: Simon Dixon

Music information retrieval (MIR) applies computing and engineering technologies to musical data to satisfy users' information needs.  This topic involves the application of artificial intelligence technologies to the processing of music, either in audio or symbolic (score, MIDI) form.  The application could be e.g. for software to enhance the listening experience, for music education, for musical practice or for the scientific study of music.  Examples of topics of particular interest are automatic transcription of multi-instrumental music, providing feedback to music learners, incorporation of musical knowledge into data-driven deep learning approaches, and tracing the transmission of musical styles, ideas or influences across time or locations.

It is intentional that this topic description is very general, but it is expected that applicants choose your own specific project within this broad area of research, according to your interests and experience. The research proposal should define the scope of the project, the relationship to the state of the art, the data and methods that you plan to use, and the expected outputs and means of evaluation.

Supervisor: Arkaitz Zubiaga

Despite the advancement in the development of LLMs, recent research has demonstrated that these models are still capable of generating texts that can be offensive to some and that others may consider to be hate speech. It is particularly the case of implicit hate speech that LLMs are still capable of generating, such as metaphorical cases where LLMs fail to identify that the underlying message contained within a text may be metaphorically offensive. Some of the angles that can be addressed in this project include to (i) investigate the shortcomings of Large Language Models (LLMs) in identifying and tackling online hate speech, studying the potential that LLMs generate samples of hate speech (e.g. via red teaming), and critical evaluation of LLMs dealing with hate speech, and (ii) to develop effective methods to safeguard LLMs from generating hate speech content, studying and furthering state-of-the-art approaches to prevent LLMs from augmenting hate speech content.

The proposed PhD programme is aligned with the EU Horizon research project ""PROCLAIM -- Polarization, Radicalization and extremism Online identifying Covert hate speech in state-of-the-art Artificial Intelligence technologies and social Media"". While aligning with the proposed idea above, the studentship is otherwise flexible and if you are interested in applying you are welcome to propose and develop your own ideas.

Supervisors: Riccardo Degl'Innocenti and Yang Hao

In this project the candidate will investigate the design of novel graphene-based, ultrafast optoelectronic devices for the active modulation of Terahertz radiation, e.g. beam steering/focusing.  The ideal candidate will be using both forward design with standard finite element methods, e.g. CST or Comsol Multiphysics, and inverse design via advanced techniques. The final fabricated devices will be then implemented for non-destructive testing, and material investigation, e.g. of ion batteries and photovoltaic materials.

 

Supervisor: Ying He

This PhD will investigate how risks arise and propagate when human work with AI systems to make decisions and coordinate tasks. It will bring together human reliability analysis, AI reasoning and socio-technical risk modelling to understand how human and AI contributions jointly influence outcomes. 

This project will assess how incomplete context, unsupported AI-generated claims and uncertainty in model outputs affect decisions.  It combines analysis of AI outputs and interaction logs with controlled user studies and computational modelling. This project will develop methods to represent dependencies between evidence, AI recommendations, human judgements and resulting actions. These models will support risk assessment by identifying where errors accumulate, which decisions have the greatest downstream consequences and where intervention may be most effective. 

Building on these findings, the project will develop and evaluate risk mitigation measures for AI-supported decision-making.  To assess their effectiveness, the project will measure decision accuracy, error recovery, workload and task completion under different interaction conditions. Expected contributions include a model of risk propagation in human-AI systems, methods for assessing decision reliability, and evaluated AI-supported interventions for risk mitigation.

Supervisors: Changjae Oh and Ziquan Lu

Robots often rely on pre-collected demonstrations and fixed models, limiting their ability to adapt to new tasks and environments. This project will investigate how robots can learn from their own successes and failures with limited human supervision. Possible directions include exploring vision-language models to understand scenes and evaluate action outcomes, and world models to predict the effects of robot actions. These models will help robots plan actions and learn from experience. The methods will be tested on robot manipulation tasks in a real-robot environment. The student will gain expertise in computer vision, multimodal learning and robot learning.

Supervisor: Dr Evangelia Kyrimi

Large Language Models (LLMs) are transforming AI, but can they reliably reason about cause and effect rather than relying on patterns in their training data? This distinction matters when asking what would happen after an intervention, or what might have happened if a different decision had been made, especially in high-stakes areas such as healthcare.

Recent work has shown promising progress in LLM causal reasoning, yet challenges remain. Models can struggle with unfamiliar causal structures, generate plausible but unjustified conclusions, and fail to recognise when the available evidence cannot answer a causal question.

This PhD will investigate how principles from causal AI and structured causal knowledge can help LLMs reason more reliably and generalise beyond familiar problems. The project will explore new approaches to causally grounded reasoning, with scope for original methodological contributions in machine learning and trustworthy AI. A key question is how formal causal knowledge can guide, rather than merely evaluate, LLM reasoning.

Healthcare will be an important application area, while the methods will aim for wider applicability. Applicants should have a strong background in computer science, AI, or machine learning, good programming skills, and an interest in LLMs or causal AI. Prior healthcare or causal-inference experience is not required.

Supervisors: Ioannis Patras and Georgios Tzimiropoulos/Dimitrios Kollias

Current generative models excel at short term generations but struggle with long-horizon planning and robust value alignment. This work will investigate structured planning algorithms with advanced alignment techniques, ensuring agents execute complex, multi-step actions that remain highly faithful to human intent. We will design planning modules that will enable models to construct, evaluate, and iteratively refine hierarchical plans prior to execution. By separating reasoning from generation, this approach mitigates hallucination and logical drift during sequential decision-making. In addition, we will investigate trajectory-level alignment. Existing RLHF models optimize for short-term proxies, ignoring long-term strategic coherence. We will explore reward shaping mechanisms that evaluate complete action sequences, ensuring all intermediate steps adhere to safety constraints.

Supervisors: Edon Kelmendi

Verification is a subfield of computer science concerned with correctness of various systems, ranging from software, to AI models, to cyber-physical systems. One aims to develop techniques and algorithms for verifying, or proving, the system behaves in the way that it was intended. This is done by studying formal mathematical models that capture the system's behaviour. 

Verification and adjacent endeavours are becoming increasingly important for multiple reasons: one, software has deeply proliferated our lives, two, more and more this software is being generated by AI models, which in turn can easily find security vulnerabilities, and three, alignment issues with the AI systems themselves. 

This project is about studying a fundamental family of such mathematical models: linear systems. They come in a few flavours: continuous or discrete, with or without conditionals, with or without bounded dimension etc. The main objective will be to develop techniques and algorithms for automatically verifying that a given linear system has a given property. 

As this project is in theoretical computer science, a strong background in mathematics is required. Particularly in algebra, formal languages, logic, and algorithms. Even more importantly, a wish to learn about these fields is paramount. 

To read more about my work, and get a glimpse of what this project is about please take a look here: https://edon.github.io/publications_theme.html . 

Supervisor: Raymond Hu

This project brings together several hot areas of research.

Modern AI and LLMs depend on protocols for coordinating task planning, reasoning steps and tool actions.  Multiagent systems (MAS) further depend on protocols to coordinate multiagent interactions (e.g., LangGraph).  The effectiveness of agentic AI depends crucially on correct protocol design and agent compliance.

Multiparty session types (MST) is an approach from programming languages theory (PL) that offers techniques for formally specifying, programming and verifying MAS protocols.  MST can be used to support verified distribution and failure handling, and prove critical properties such as freedom from communication mismatches and deadlocks.

The PhD student will work in the new and exciting intersection of these areas.  We can discuss the exact topics: we have freedom to go more into the AI side or the MST/PL side; more into the theory or the practical side.  Example topics:

* Develop an MST-based language for safe coordination of AI agents and agentic LLM.
* Combine MST with logical refinements to specify and verify tree-of-thought protocols.

The student will join the research team of Dr Raymond Hu.  My other projects include working on MST with industry partners Monzo Bank and SAP in our UKRI funded research project (£646k), DymSUM.

The student will also join the Centre for Fundamentals of AI and Computational Theory (FACT).

Contact Dr Raymond Hu (r.hu@qmul.ac.uk) ASAP to discuss the project.

Supervisor: George Fazekas

Learning meaningful representations of music and audio is central to advances in music understanding, retrieval, recommendation and generation. Recent progress in self-supervised and multimodal learning, large-scale pretrained models and generative AI has opened exciting possibilities, yet important challenges remain around musical structure, generalisation, interpretability, controllability and the use of limited or imperfectly annotated data.

This PhD will investigate novel approaches to music and audio representation learning, exploring how data-driven methods can be enriched by musical knowledge, structural priors, or insights from signal processing. Potential directions include integrating audio and symbolic representations, learning hierarchical and relational musical structures, exploiting text-audio and other multimodal models, developing interpretable and controllable representations, and connecting representation learning with generative modelling. The project offers scope to challenge existing paradigms and develop new methods that advance the state of the art in music and audio AI.

The research should also consider the broader implications of these technologies, including dataset bias, transparency, responsible use of training data and/or attribution of creative contributions. Candidates will have the opportunity to shape the research direction around their interests and expertise.

Supervisor: Emmanouil Benetos

The field of music information retrieval (MIR) has historically focused on audio and machine-readable scores as its core representations and modalities under investigation. However, music as an art form is inherently multimodal: ways to capture music performances and music compositions include but are not limited to video, images (from music scores to album art), lyrics and text descriptions in addition to audio. This PhD project will investigate and propose new methods from the field of Multimodal AI towards creating and evaluating multimodal foundation models for music understanding. Part of the project will involve data collection towards a multimodal MIR corpus, as well as benchmarking existing models and methods against multimodal MIR tasks. The successful candidate will investigate, propose and develop novel machine learning methods for MIR across multiple modalities, resulting in models that can accelerate and broaden MIR research but also advance the broader field of Multimodal AI.

 

Supervisor: Yang Hao

This project will develop flexible, skin-conformal electromagnetic (EM) sensors for non-invasive, continuous monitoring of cardiovascular and respiratory signals. Built on stretchable antenna and metasurface elements printed on thin polymer substrates, the sensors will couple wirelessly to a low-power readout unit, removing the need for rigid electrodes or wired connections. The student will design and fabricate the flexible EM sensing elements, characterise how they respond to tissue loading and motion, and build an on-device signal-processing pipeline to separate genuine physiological signal from motion and coupling artefacts. The work sits within the antennas and electromagnetics group at Queen Mary University of London, alongside ongoing projects on wearable cardiovascular sensing, software-defined RF materials and neuromorphic edge inference, and will be jointly supervised with a partner institution in China under the CSC scheme.
We are looking for a candidate with a strong background in electromagnetics, RF/microwave engineering, materials science or a closely related discipline, and an interest in wearable health technology.

Supervisor: Kamyar Mehran

This fully funded China Scholarship Council (CSC) PhD project will develop a communication-efficient, AI-based control architecture for distributed DC microgrids. DC microgrids are increasingly attractive for remote communities and renewable-rich networks. They depend on large numbers of power electronic converters and on a communication layer that must stay light, scalable and reliable. The project will design distributed, AI-based control and coordination strategies that cut communication load while keeping the bus voltage stable and the power shared across heterogeneous distributed generation sources and energy storage. In parallel, the project will build embedded condition-monitoring methods that estimate the health of power electronic converters. The goal is to detect degradation early, with the monitoring informing the control layer so that the microgrid can derate, re-share load or schedule maintenance before a fault. The work will be validated through RPCS laboratory’s microgrid test bench, with opportunities to engage with industry partners. Applicants should have a background in power electronics, control systems or embedded systems, and knowledge of real-time simulation and hardware-in-the-loop testing.

Supervisor: Haim Dubossarsky

"Can we learn something fundamentally new about language, knowledge, or human culture by studying artificial intelligence? Rather than treating AI models simply as tools for solving problems, this project explores how their mathematical structure and behaviour patterns can become instruments of scientific discovery.

The goal of this PhD is to develop mathematical and/or behavioural approaches to understanding how LLMs represent and process information, and use these insights to investigate questions beyond AI on the one hand and human language and culture on the other hand. 

Methods: Mathematical approaches such as topology (TDA), geometry, information theory, and mechanistic interpretability can illuminate the internal structure of neural networks. Behavioural experiments inspired by psychology can probe model representations, capabilities, and reasoning. Both types of methods are useful outcomes of this research.

Applications: Interpretability methods can help study broad questions of language and culture from historical language change, linguistic diversity, literature, translation, and the human mind. These challenges can, in turn, inspire new approaches to understanding AI.

Applicants from computer science, mathematics, cognitive science and experimental psychology, computational linguistics and NLP, and related fields are welcome. Research questions and approaches will reflect the student's interests and expertise."

Supervisor: Dimitrios Kollias

This research aims to develop next-generation AI models and reasoning frameworks for sustainable edge computing. With the rapid growth of intelligent sensors, mobile devices and IoT systems, there is an urgent need for efficient, interpretable, energy-aware AI that can operate directly on edge devices, reducing latency, bandwidth use, environmental impact.

The project will focus on designing lightweight, adaptive, explainable AI algorithms capable of performing complex reasoning tasks in resource-constrained environments. It will explore:

(i) Sustainable AI design, including model compression, neural architecture search, green optimization to minimize computational cost and carbon footprint;

(ii) Edge intelligence, enabling distributed learning and on-device inference through federated, collaborative, privacy-preserving AI paradigms.

(iii) Neuro-symbolic reasoning, integrating deep learning with logical and causal reasoning to enhance interpretability, robustness, trustworthiness.

(iv) Real-world applications, focusing on behaviour understanding/analysis, smart healthcare, intelligent environments, human-centred analytics, where energy-efficient and privacy-preserving AI systems can deliver measurable social and psychological impact.

By advancing the foundations of sustainable edge AI & reasoning, this project aligns with the global vision for green, responsible, human-centred AI. The research combines innovation across AI, sustainable computing and cognitive reasoning.

Supervisor: Maria Likata

As AI agents become part of everyday life, their usefulness increasingly depends on sustaining interactions over days, weeks or months. Such interactions require systems to track changing contexts, needs and preferences, combining information across time and modalities to support coherent, personalised responses. Evaluating these capabilities is challenging because interaction quality depends on the evolving relationship and conversational history, and there is rarely a single correct response. This motivates research into hybrid temporal architectures, transparent reference-free evaluation and explainable generation, alongside methods that protect sensitive information and reduce bias as agents accumulate and synthesise personal context.

Prof Liakata has a position for a CSC scholar who will undertake a PhD project at the intersection of natural language processing (NLP), machine learning (ML), multimodal AI and responsible innovation. The successful applicant will be expected to work on one or two of the following topics during their PhD: 
Hybrid temporal architectures for improving personalisation and generation in long-horizon interactions
Transparent, reference free evaluation of multi-modal human-agent interactions
Methods for privacy preservation and bias reduction (including synthetic data) in generation for dialogue systems and information synthesis
Explainable  factually-grounded generation with LLMs

The project will be closely linked to the work within the AdSoLve project: https://adsolve.github.io/.

For queries please get in touch with Prof Liakata (m.liakata@qmul.ac.uk)

Back to top