Imaging, Data and Communications
Continuous wearable sensing can flag postoperative deterioration earlier than intermittent observations, but prediction alone does not quantify the impact of alternative actions. Using the EMUs multi-country “shadowmode” datasets linking Sibel ANNE® One waveforms to time-stamped events and 30-day outcomes (NCT06565559), this PhD will develop AI-first causal models: self-supervised representations of physiology, counterfactual sequence prediction, and off-policy evaluation of escalation strategies. The goal is decision support that surfaces expected benefits, harms, and uncertainty across settings, with robustness and equity tests for global surgery.
Why study with us What is CHAI?
CHAI is the Causality in Healthcare AI Hub that unites an international consortium of universities, industry partners, government bodies, and regulatory entities to develop cutting-edge causal AI innovations to enhance patient care and outcomes. We are working to develop an explainable causal AI platform specifically addressing unique challenges from healthcare across prevention, diagnosis, and treatment. We believe that to move AI forward, we need to build models incorporating both theory and observation. This will create models that are more transportable, explainable, fair, and of more direct relevance for decision support. The CHAI Hub co-develops research with clinical experts, policy makers, and patients to address complex and heterogenous data structures, nuanced real-world problems, and a rapid pathway to societal and economic impact.
What is Sibel Health Inc.?
Sibel Health is a healthcare technology company that develops advanced wireless wearable sensors to continuously monitor vital signs such as heart rate, respiratory rate, temperature, and movement. Its solutions are used in hospitals, clinical research, and remote patient monitoring to provide accurate, real-time health data. The company’s mission is to deliver “Better Health Data for All®” by improving the quality, accessibility, and reliability of patient data so clinicians and researchers can make more informed decisions and ultimately improve patient outcomes worldwide. What can you get from working with us? At CHAI, we’re building a dynamic team of innovators, researchers, and professionals dedicated to transforming healthcare through causal AI. If you’re passionate about making a meaningful impact in healthcare, AI, and data science, CHAI offers a unique opportunity to work at the forefront of causal AI development with a focus on real-world solutions and societal benefit. We offer:
- An innovative environment. Work on cutting-edge AI solutions for healthcare challenges, from optimising treatment outcomes to advancing diagnostic capabilities.
- A collaborative culture. Join a multidisciplinary team of experts across AI, healthcare, and data science, with collaboration at the heart of everything we do.
- Professional growth. We offer opportunities for professional development, including training, mentorship, and access to leading research.
- Commitment to sustainability. At CHAI, we integrate environmental responsibility into our projects and strive to make sustainable choices in AI innovation. View our website at this link to see what it is like to be an early career researcher with us: https://www.chai.ac.uk/being-a-chai-ecr
References
https://clinicaltrials.gov/study/NCT06565559 https://bmjopen.bmj.com/content/bmjopen/15/10/e104463.full.pdf
Project background
Postoperative deterioration is often detected late because ward monitoring is intermittent and early warning scores compress complex physiology into sparse snapshots. Continuous wearable waveforms create an opportunity to learn patientspecific trajectories and early phenotypes of complications, but real impact requires causal, action-aware modelling rather than “higher AUC”. EMUs (NCT06565559) is an international, prospective, observational cohort collecting up to 10 days of continuous chest/limb sensor data in “shadow mode” alongside standard clinical data and 30-day outcomes. Sibel Health develops wireless, clinical-grade continuous monitoring (ANNE® One) capturing ECG waveforms and vital-sign streams suitable for hospital and resource-variable environments.
The CHAI Hub focuses on explainable, transportable causal AI platforms that integrate theory and observation to support decisions. This PhD will fuse these strengths to estimate how different postoperative pathways (e.g., earlier escalation vs watchful waiting) change outcomes, and how those effects generalise across health systems.
Research aims
1. Learn high-fidelity, clinically grounded representations of postoperative physiology from waveforms and context.
2. Estimate heterogeneous causal effects of time-varying decisions (escalation, antibiotics, imaging, ICU transfer) using AI-augmented counterfactual modelling and off-policy evaluation.
3. Develop transportable and fair causal decision policies that remain reliable under domain shift, missingness, and resource constraints, and present trade-offs (benefit/harms/uncertainty) for shared clinical decision making.
Data & methodology
Data will extend EMUs by linking ANNE® One waveforms and derived features to time-stamped actions, observations, complications and 30-day outcomes. Methods will emphasise AI-centric causal learning: self-supervised pretraining (contrastive/masked modelling) on waveforms; transformer/state-space models for latent physiologic state; neural causal effect estimators (representation-balancing + doubly robust learners) for static and sequential treatments; and causal reinforcement learning for dynamic treatment regimes with off-policy evaluation and calibrated uncertainty. Robustness will be tested via invariant/causal representation learning and fairness audits across sites.
Expected outcome and impact
Outputs of the work as a whole include:
(i) a pretrained physiologic foundation model for postoperative monitoring;
(ii) validated counterfactual estimators for action timing and escalation thresholds, with subgroup and site heterogeneity;
(iii) an interpretable prototype that compares “what if” trajectories under alternative care pathways and highlights trade-offs and uncertainty;
(iv) evidence on transportability and equity of causal policies across high- and lower-resource settings; and
(v) trial-ready specifications for an interventional evaluation of AI-guided monitoring strategies.
Timescale of expected outcomes
Year 1: governance, waveform pipeline, representation pretraining, and clinical target-trial specification using EMUs extensions.
Year 2: build and validate counterfactual sequence models and neural effect estimators; define candidate policies and off-policy evaluation plan.
Year 3: transportability, robustness, and fairness analyses across countries; prototype trade-off interface with clinical co-design.
Year 4: external validation on newly accrued data; ablation and shift-stress tests; publications, thesis, and a pragmatic trial protocol with deployment metrics. Student training and development The student will gain depth in modern causal ML and sequential decision making (causal representation learning, off-policy evaluation, causal RL), plus signal processing and self-supervised modelling of physiological waveforms. Training will be supported through CHAI Hub and Sibel Health.
How to apply
As part of your application, include the name of Professor Sotirios Tsaftaris in the statement, and discuss your motivation for doing a PhD, what attracts you for this PhD position (including the group, the university, CHAI, and Sibel Health) and what your aspirations are for after the PhD. In the application title, add '“Project funding by Prof. Tsaftaris and Sibel Health for CHAI”. Please also make sure to include your most up-to-date CV as part of the application submission.
For frequently asked questions regarding applying for a PhD position in Causal AI in understanding medical images, please visit here: https://vios.science/faq/
Minimum entry requirements
- a 2:1 undergraduate degree (or equivalent).
- the University’s English language requirements.
What we are looking for?
We encourage applicants to provide evidence of the qualities below in their application, through coursework, projects, work experience, or independent learning. You do not need direct experience within every area, but you should show motivation to learn, awareness of why these skills matter, and a thoughtful attitude toward developing them during the programme.
We recognise that applicants come from many different starting points. If you have taken a non-traditional route or faced circumstances affecting your performance, we encourage you to describe this in your application so we can consider it appropriately.
A successful candidate will:
- Be a student home to the UK who will be based in Edinburgh;
- Have a strong foundation in either computer science, AI, cognitive science, mathematics, physics, engineering, biomedical science, biological science, or clinical & public health sciences;
- Be able to demonstrate skills training in programming, data analysis, or computational thinking, and ideally evidence of successful deployment of these skills in the form of a project;
- Have a genuine interest in biomedical or health applications and awareness of the challenges specific to this sector;
- Appreciate that AI in medicine.
This project is fully funded for a home student (fees and stipend).
To qualify as a Home student, you must fulfil one of the following criteria:
- You are a UK student OR
You are an EU student with settled/pre-settled status who also has 3 years residency in the UK/EEA/Gibraltar/Switzerland immediately before the start of your programme.
International students are not eligible.
Objectives:
- Develop new adaptive data-driven sensor tasking and management strategies.
- Explore combination of Bayesian signal processing and modern machine learning solutions.
- Implement and evaluate developed techniques in simulated and real world scenarios.
Description:
Sensor networks, sensor fusion and management techniques address key challenges in intelligence, surveillance, target acquisition, and reconnaissance (ISTAR). Opportunities in adaptive data-driven sensor tasking and resource management include adaptive sensor placement, adaptive waveform design to reflect the target reflection characteristics and channel environments, and adaptive sensor selection. Although these problems have solutions in specific use cases, this project will consider scenarios with broader applications involving multiple heterogeneous sensors on single or multiple airborne autonomous collaborative platforms (ACPs).
The solutions developed in this should be robust to dynamic and congested environments, adverse weather conditions, and mutual sensor interference. A range of algorithmic and signal processing or machine learning technologies will be considered, as well as specific technical challenges. For example, this project could consider aspects related to wide area motion imaging (WAMI), position, navigation, and timing issues (PNT); robustness to adversarial attack; sensor fusion and tracking applications; use of machine learning and Monte Carlo methods. Probabilistic and Bayesian frameworks will be preferred to enable uncertainty quantification and management.
This project will be jointly supervised by Professor James Hopgood and Professor Mike Davies.
Research Theme
Industry Partner
Leonardo (as part of the Prosperity Partnership Smart Products Made Smarter)
Applications
First-round applications have closed and the applications for SPADS are now being considered on a gathered field basis, where applications will be considered at the end of every month until all places are filled.
UK 2:1. in computer science / computer systems or MSc in geographical information science.
Minimum entry qualification - an Honours degree at 2:1 or above (or International equivalent) in a relevant science or engineering discipline, possibly supported by an MSc Degree.
Further information on English language requirements for EU/Overseas applicants.
This PhD project aims to develop Physics-Informed Neural Networks (PINNs) for the prognosis and understanding of instabilities in zero-carbon reactive flows, with a particular focus on hydrogen and/or ammonia combustion. The project will integrate advanced machine learning techniques with laser imaging to create predictive, data-efficient models capable of capturing the complex, multi-scale dynamics of reacting flows. This project is under the Royal Society International Exchange Programme in collaboration with Shanghai Jiao Tong University.
Imaging methods such as absorption or emission tomography provide rich, high-resolution spatio-temporal data on flow characteristics. These measurements will be combined with governing physical laws embedded within PINN frameworks to infer hidden flow states, identify instability precursors, and forecast the onset and evolution of reactive-flow instabilities.
The research will involve developing tailored PINN architectures for reactive flows, designing strategies to assimilate experimental laser-imaging data, and validating models against laboratory-scale combustion experiments. Emphasis will be placed on uncertainty quantification, robustness to sparse and noisy data, and real-time or near-real-time predictive capability.
The project is highly interdisciplinary, spanning laser imaging, applied mathematics, and machine learning, and is suited to candidates with strong interests in both imaging and data-driven methods.
Primary objectives:
- Develop PINN frameworks tailored to zero-carbon reactive-flow systems.
- Integrate imaging diagnostics with PINNs for model training.
- Identify and characterise precursors to instabilities from spatio-temporal flow fields.
- Create predictive models for early-warning prognosis of instability.
- Validate PINN predictions against controlled laboratory combustion experiments.
- Investigate uncertainty quantification and robustness.
Required skills:
1. Matlab/Python/Tensor-flow coding (mandatory)
2. Machine learning algorithms and coding (mandatory)
3. Tomography and inverse problems
4. Flow-field mechanics and dynamics
5. Optical/electronic experimental skills.
Please note that this advert will close as soon as a suitable candidate is found.
Applicants should have an Undergraduate degree in Electronic and Computer Science or Mechanical Engineering, possibly supported by an MSc Degree
Please also refer to the University’s English language requirements.
Applications are welcomed from self-funded students, or students who are applying for scholarships from the University of Edinburgh or elsewhere
We are seeking an outstanding Doctoral Candidate (DC) for a fully funded studentship as part of the prestigious Horizon Europe MSCA Doctoral Network “Advanced Network Connectivity using Harmonious Optical and Radio Technologies (ANCHOR)” The studentship covers UK fees and provides a generous salary, consumables, and travel budget.
Extending high-speed wireless data transfer to the underwater environment has several benefits that include environmental protection, national security, mineral exploration, and monitoring of vital underwater infrastructure among others. This need has led to increased research in underwater optical wireless communication (UOWC). Recent studies in controlled in-lab testbeds have shown the possibility of a Gbps data rate. But achieving reliable, robust and resilient data transfer in the challenging underwater channel remains a challenge. This project will study UOWC and investigate approaches to link the underwater network to the much wider land and space communication networks.
This project is also to study the existing radiative transfer theoretical framework of photon propagation in underwater to formulate a signalling/waveform design that is best suited to the UOWC channel. This will be followed by designing a new resilient waveform for UOWC and developing a computer-based simulation to evaluate its performance. Finally, the study will use existing UOWC channel emulator to demonstrate the feasibility of the new signalling/waveform.
The ideal candidate for this project would have first degree and MSc/MEng in physics, electronic and electrical engineering or a related discipline. A passion for experimental work and capability to develop/apply mathematical modelling techniques will be particularly important for this project.
Expected start date: June 2026 or shortly thereafter.
Undergraduate and postgraduate degrees in Physics, Electronic and Electrical Engineering or related discipline.
Tuition fees + stipend are available for Home and International students.
To qualify for the funding, the applicant must not have resided or carried out their main activity (work, studies, etc.) in the UK for more than 12 months in the past 36 months.
Research Theme
Multi-Agent Systems and Data Intelligence
Aim
Develop a framework for the certification of AI-based systems for safety and security.
Objectives
1. Formalise existing methods for assessing AI-based systems to identify building blocks for the framework.2. Develop analytic methods combining compositional reasoning and machine learning pipelines to evaluate AI-based systems.3. Apply the approach to the context of remote attestation in multi-agent systems.
Description
This project targets the fundamental question of the certification of AI-based data-systems for security and safety. This has been identified as a crucial open question by the 2023 AI Safety Summit and its recent follow-ups. Challenges posed by novel AI models require to re-think the analysis and certification of systems. By getting inspiration from the analysis of security and safety of software systems, this project combines knowledge and practice of machine learning and AI, and formal approaches of compositional reasoning for verification. The project is to conduct experiments targeting the problem of forming coalitions of agents in a multi-agent system where some of the agents do not belong or have been compromised. A solution requires remote attestation, whereby an agent proves that they are uncompromised (aside from authenticating themselves). The project contributes to Dstl’s secure by design adoption by providing a means to construct and evaluate secure AI-based systems.
Bengio et al. 2025. The Singapore Consensus on Global AI Safety Research Priorities. https://doi.org/10.48550/arXiv.2506.20702 Dalrymple et al. 2024. Towards Guaranteed Safe AI: A Framework for Ensuring Robust and Reliable AI Systems. https://doi.org/10.48550/arXiv.2405.06624 Dstl. 2025. Secure by Design Problem Book. https://www.gov.uk/government/publications/secure-by-design-problem-book
Applications
First-round applications have closed and the applications for SPADS are now being considered on a gathered field basis, where applications will be considered at the end of every month until all places are filled.
Minimum entry qualification - an Honours degree at 2:1 or above (or International equivalent) in a relevant science or engineering discipline, possibly supported by an MSc Degree.
Further information on English language requirements for EU/Overseas applicants.
Home fee rate and stipend available for this position.
Research Themes
Aim
Optimise PNT as a Service using satellite signals of opportunity (SoOP) across LEO/MEO/GEO.
Objectives
- Establish a testbed that will enable PNT services using satellite SoOP.
- Benchmark the performance of PNT services using satellite SoOP against GNSS-based systems.
- Optimise data management of a the system for performance and robustness
- Develop AI-driven models for the selection of the optimum SoOP to be used in a PNT as a Service scenario
Description
GNSS-enabled positioning, navigation, and timing (PNT) is under increasing threats from interference and spoofing.
Satcom systems offer signals of opportunity (SoOP) that can be used to deliver PNT services. They come from an abundant number of satellites across LEO/MEO/GEO, each offering distinct advantages in terms of signal strength or coverage. SoOPs are also available across a wide range of frequencies (from C-band to Ka-band) and multiple angular directions, thus making SoOPs a highly reliable option of pervasive PNT services.
This project will access a UK-wide network of monitoring and reference stations publishing SoOPs across C-, Ku- and Ka-bands. Data from this network will be used in conjunction with signals captured at a user terminal to deliver and benchmark PNT. The data management aspects of such a system will be investigated to optimise performance. AI-driven algorithms for selecting optimum usage of the proposed system will be developed and evaluated.
Industry Partner
This project will be in collaboration with PNTaas
Applications
First-round applications have closed and the applications for SPADS are now being considered on a gathered field basis, where applications will be considered at the end of every month until all places are filled.
Minimum entry qualification - an Honours degree at 2:1 or above (or International equivalent) in a relevant science or engineering discipline, possibly supported by an MSc Degree.
Further information on English language requirements for EU/Overseas applicants.
Full funding is available for this position.
Dynamical system models have been the main pillar of conventional model-based approaches in control, signal processing and sensor fusion: Sensor signal processing and inference algorithms for applications such as multi-object detection and tracking, robotic simultaneous localisation and tracking (SLAM) and calibration of autonomous networked sensors are designed by combining the known physics and stochastic elements into dynamic system models. Model inaccuracies can be mitigated to achieve significant performance gains in inference and decision-making by leveraging data and model size, following the recent advances in machine learning. However, learning from data in dynamical system models to jointly address the epistemic and aleatoric uncertainties involved remains a challenge stemming from a range of factors such as noisy data and labelling, inhomogeneous sampling, model complexity and the intractability of posterior inference.
This PhD project aims to conduct underpinning research to address problems involving the learning of hierarchical, time-varying multi-dimensional state space models for dynamic objects/phenomena, noisy measurements made in complex backgrounds and/or in the presence of calibration errors.
The incumbent will have the opportunity to steer the direction of the research in consideration of the impact on engineering problems, including learning models for complex backgrounds in radar detection, learning of birth and trajectory models to improve detection and tracking, or semi-supervised/unsupervised training of sensor data classifiers.
The application documents should provide sufficient information and evidence regarding the applicant’s background and goals, and include:
- A Personal Statement
- CV
- Official certificates and transcripts (and English language translations, if applicable)
- English Language certificate (if applicable, for more information, please see here: https://study.ed.ac.uk/programmes/postgraduate-research/947-engineering)
- 2 Recommendation Letters
- Research proposal (optional for candidates who are not self-funded or applying for scholarships. Must be related to the project description)
Dr M Uney is an experienced scientist with interests in signal processing, machine learning, probabilistic models and Bayesian computations in sensor fusion and signal & information processing.
Prof Mike Davies holds the Jeffrey Collins Chair in Signal and Image Processing at the University of Edinburgh. His research interests are in machine imaging, data-driven computational sensing and imaging, and sensor and information fusion.
This position is open to UK/EU nationals and international applicants, and offers an annual stipend of £21,935 (as of 2024-25 fiscal year, subject to revision) for a duration of 3.5 years with £5000 research expense funds for the duration of the study.
UK nationals and other eligible applicants might choose to align this studentship with the Sensing, Processing and AI for Defence and Security Centre for Doctoral Training (SPADS CDT) at the University of Edinburgh’s School of Engineering, and benefit from an enhanced stipend (subject to CDT administration approval and security clearance, see the eligibility criteria in the link) upon acceptance. Eligible applicants should state their interest in their Personal Statement.
Minimum criteria:
- A 2:1 Master’s Degree in engineering, computer science, statistics, physics or in an appropriate subject or a related discipline.
- The University’s English language requirements
Preferable criteria:
- A 1st class Undergraduate Degree in engineering, computer science, statistics, physics or in an appropriate subject or a related discipline.
- A 1st class Master’s Degree in engineering, computer science, statistics, physics or in an appropriate subject or a related discipline.
- Experience in scholarly writing.
Tuition fees + stipend are available for Home/EU and International students
Applications are welcomed from self-funded students, or students who are applying for scholarships from the University of Edinburgh or elsewhere