At a Glance
- Tasks: Develop and evaluate methods for knowledge editing in large language models.
- Company: Join a leading research team at UEA focused on AI advancements.
- Benefits: Receive a competitive stipend, tuition coverage, and access to high-performance computing resources.
- Other info: Opportunity to work with cutting-edge technology and contribute to innovative AI applications.
- Why this job: Make a real impact on AI by improving the reliability and accountability of language models.
- Qualifications: Must have a 2:1 degree in Computer Science.
The predicted salary is between 21805 - 21805 £ per year.
About the Project
Emerging techniques in model editing, also discussed in the literature as knowledge editing and memory editing, offer promising approaches for correcting erroneous or harmful information, incorporating specialised knowledge, and improving the usability, security, and accountability of large language models (LLMs). The overall aim of this PhD project is to develop and evaluate methods for reliable, localised, and generalisable knowledge updating in large language models. Using accessible open-weight language models suited to available computational resources, including NVIDIA RTX 6000 Ada 48GB GPUs and UEA high-performance computing (HPC) facilities, you will adapt established editing methods, such as ROME and MEMIT, and compare them with retrieval-augmented generation (RAG), external memory mechanisms, and conventional fine-tuning.
Experiments will investigate how models can incorporate new information or knowledge, correct inaccurate information, and remove outdated or unwanted information while minimising unintended changes to existing capabilities. Evaluation will use established factual model-editing benchmarks, such as CounterFact and zsRE, supplemented where appropriate by a domain-specific dataset. Performance will be assessed in terms of editing efficacy (whether the target knowledge is successfully updated), generalisation (whether updates remain effective across paraphrases and related queries), locality (whether unrelated knowledge and capabilities are preserved), and retention (whether updates persist following subsequent edits or conflicting information). Further evaluations will examine factual consistency, general reasoning, and robustness to adversarial prompts. Where feasible, the project will also investigate knowledge provenance and update traceability to support accountability. Comparative experiments will establish the strengths, limitations, and trade-offs of different knowledge-updating approaches.
You will also have the opportunity to deploy selected models in embodied AI systems or social robots, such as Furhat, and conduct human-AI interaction experiments to investigate how knowledge updates affect responses, consistency, and behaviour during real-time interactions. The project will contribute practical methods and empirical evidence for developing reliable, trustworthy, and accountable LLMs for knowledge-intensive applications.
Entry Requirements
The standard minimum entry requirement is 2:1 in Computer Science.
Mode of Study
Full-time
Start Date
1 October 2027
Additional Funding Information
This PhD project is in a competition for a Faculty of Science funded studentship. Funding is available to UK applicants and comprises ‘home’ tuition fees and an annual stipend for 3 years. The current UEA / institutional PhD stipend rate is £21,805 and will be subject to any agreed inflationary uplift for 2027/28.
PhD Studentship: Reliable and Generalisable Knowledge Editing in Large Language Models in Norwich employer: UAG
Queen Mary University of London is an exceptional employer, offering a vibrant work culture that fosters academic excellence and innovation. With opportunities for professional development and the unique experience of teaching in both London and Hainan, China, employees are encouraged to grow their skills while making a meaningful impact on students' academic journeys.
StudySmarter Expert Advice🤫
We think this is how you could land PhD Studentship: Reliable and Generalisable Knowledge Editing in Large Language Models in Norwich
✨Get Involved in Data Science Meetups
Tap into local data science meetups or workshops to connect with fellow enthusiasts and professionals. These events are goldmines for networking, and sometimes even lead directly to job openings at companies like UAG!
✨Show Off Your Projects
Start building a public portfolio showcasing your data science projects on platforms like GitHub or personal websites. Highlight unique analyses or models you've developed. This not only demonstrates your skills but also gets your name out there for roles like PhD Studentship: Reliable and Generalisable Knowledge Editing in Large Language Models at UAG.
✨Leverage Professional Networks
Join professional bodies related to data science, like the Data Science Society or similar organisations. Getting involved can lead to mentorship opportunities and insider knowledge about full-time positions at companies like UAG.
✨Apply Directly through Our Website
When you find a suitable opening like PhD Studentship: Reliable and Generalisable Knowledge Editing in Large Language Models at UAG, make sure to apply directly through our website. It gives you an edge and shows you're keen to join our team. Plus, who doesn’t love a direct application? It’s easier than navigating through job boards!
We think you need these skills to ace PhD Studentship: Reliable and Generalisable Knowledge Editing in Large Language Models in Norwich
Some tips for your application 🫡
Show Off Your Projects:In the world of data science, your projects can speak volumes about your skills. Make sure to showcase a few key projects in your CV or portfolio, especially those that highlight your ability to work with data sets, build models, or use relevant tools like Python, R, or SQL. Don’t forget to include links to any GitHub repositories if applicable!
Quantify Your Achievements:Employers love numbers! When drafting your CV, highlight your achievements with quantifiable results. For instance, mention how your data analysis led to a certain percentage increase in efficiency or revenue at a previous job or project. These details can really make your application pop!
Craft a Tailored Cover Letter:For a full-time role at UAG, your cover letter should reflect your passion for data science and your excitement about the specific projects or values of the company. Dive into why you’re a good fit, how your skills align with their needs, and any unique perspectives you can bring to the team.
Stand Out with Relevant Courses and Certifications:Although experience talks, relevant courses or certifications can be your ticket to impressing hiring managers at UAG. Mention any standout courses you've completed that equipped you with essential skills, such as machine learning certifications or data visualisation courses. This shows your commitment to continuously developing your skills in the field!
How to prepare for a job interview at UAG
✨Brush Up on Your Statistics
For a data science role, we need to seriously sharpen our statistics skills. Get ready to tackle technical questions on probability distributions, hypothesis testing, and regression analysis. These are often the bread and butter of data science interviews, so don't just skim over them!
✨Showcase Your Projects
Prepare a killer portfolio showcasing your data science projects. We should include details about the datasets used, the tools and techniques applied, and the impact of your findings. If we can walk them through a particularly challenging project or a cool visualisation that had real-world implications, it’ll really make us stand out!
✨Get Comfortable with Python and R
Most data science positions require us to be proficient in programming languages like Python and R. We should practice common libraries like pandas, NumPy, and scikit-learn, and be ready for live coding exercises or algorithm questions. Showing off our coding chops can really impress the interviewers at UAG!
✨Prepare for Case Studies
Expect to encounter real-world case studies during the interview. We might be asked how we’d approach a data problem or analyse a dataset to extract insights. It's essential to think out loud and demonstrate our problem-solving process so that the interviewer can see our logical thinking in action.
