- Institution
- University of East Anglia
- Department
- School of Computing Sciences
- Supervisor
- Dr Farhana Ferdousi Liza
- Location
- Norwich, United Kingdom
- Field
- Computer Science & AI › Knowledge editing in large language models
- Specialisation
- AI safety, Large language models
- Funding
- Fully funded
- Pay
- GBP 20,409 per year
- Tuition
- Covered
- Duration
- 3 years
- Start
- 1 October 2027
- Deadline
- 24 Feb 2027
- Degree required
- 2:1 in Computer Science
- Language
- Not stated
- Eligibility
- Open to UK applicants; the standard minimum entry requirement is a 2:1 in Computer Science.
Overview
The project sits in the School of Computing Sciences at the University of East Anglia, Norwich, and is supervised by Dr Farhana Ferdousi Liza, whose research covers large language models, trustworthy AI and natural language processing.
The work addresses knowledge editing: methods for correcting wrong or harmful content in large language models, adding specialised knowledge, and removing outdated information. The aim is to make such updates reliable, localised and generalisable, and to compare competing approaches on established benchmarks.
The studentship is offered through a Faculty of Science competition and is full-time, starting 1 October 2027.
Research project
The researcher will adapt established editing methods such as ROME and MEMIT and compare them with retrieval-augmented generation, external memory mechanisms and conventional fine-tuning. Experiments use accessible open-weight models and the available NVIDIA RTX 6000 Ada 48GB GPUs and UEA high-performance computing facilities.
Evaluation uses benchmarks such as CounterFact and zsRE, with a domain-specific dataset where appropriate, and covers:
- editing efficacy, generalisation across paraphrases and related queries, locality and retention after later edits
- factual consistency, general reasoning and robustness to adversarial prompts
- where feasible, knowledge provenance and update traceability
There is also the option to deploy selected models in embodied systems or social robots such as Furhat and run human-AI interaction experiments.
Funding
The project is in a competition for a Faculty of Science funded studentship. Funding is available to UK applicants and comprises home tuition fees plus an annual stipend for 3 years.
The current UEA institutional PhD stipend rate is stated as £20,409, subject to any agreed inflationary uplift for 2027/28.
Eligibility
The standard minimum entry requirement is a 2:1 in Computer Science.
Funding is available to UK applicants.
How to apply
Applications are made for this PhD project, which is entered into a competition for a Faculty of Science funded studentship. The advert gives a closing date of 24 February 2027 and a start date of 1 October 2027. The source does not describe the application steps in detail.
You apply on the university's own page. UniPositions never collects applications, and no real position charges applicants a fee. Scam warning
Before you apply
- No language requirement is stated: confirm which test or certificate is accepted.
How this compares
10 open Computer Science & AI positions in United Kingdom are listed here; 9 of them state a salary or stipend. All Computer Science & AI positions · All positions in United Kingdom
Similar open positions
PhD Studentship: Self-Supervised Learning for Sparse and Heterogeneous Spatio-Temporal Data
PhD project at UEA developing self-supervised learning methods for sparse, irregular and heterogeneous spatial and temporal data, starting from an environmental application predicting microbial contamination in bathing waters. Suits a graduate in computing or data science who wants to build new mach
PhD Studentship: Digital Twins for waste reduction in Industry 5.0
A PhD project developing AI- and IoT-enabled digital twin frameworks that monitor manufacturing processes, detect anomalies and defects, and help predict and reduce industrial waste, in line with Industry 5.0. Suited to graduates in computing, engineering or a related area with an interest in machin
PhD Studentship: Quantum-Resilient Security for Safety-Critical Vehicular Systems
A PhD project developing post-quantum cryptographic security for connected and autonomous vehicles in 6G networks, weighing security strength against computation and latency costs. Suits graduates with strong programming or networking skills and an interest in cybersecurity and applied cryptography.
PhD Studentship: Distributed Machine Learning over Mobile Networks
PhD on AI over radio access networks: optimising distributed and mobile inference of large language models by adapting token, memory and bandwidth use in real time, with the aim of informing 6G network design. Suits graduates in computer science, statistics or electronic/information engineering with
PhD Studentship: Smart Optical Design: AI Tools for the Next Generation of Cinematic Camera Lenses
PhD project at the University of Surrey building AI and differentiable-rendering tools that simulate and inversely design cinematic camera lenses. Suits graduates in computer science, electronic engineering, maths or physics with strong maths and Python/C++ machine-learning skills.
PhD programme in Machine Learning Systems
A cohort-based PhD programme in Machine Learning Systems at the University of Edinburgh, combining computer systems, machine learning, hardware, sensors and control, with company internships and entrepreneurship training. Suits graduates with strong computing and mathematical background and programm
Source
This listing was written from University of East Anglia (official page), last checked 17 Nov 2026. We read the page and wrote the facts in our own words; always confirm details on the official page.



