Enhancing Large Language Models Reasoning via Computational Argumentation.

We are recruiting new Doctoral Researchers to our EPSRC funded Doctoral Training Partnership (DTP) PhD studentships starting 1 October 2024. Applications are invited for the project title Enhancing Large Language Models Reasoning via Computational Argumentation.

Successful applicants will receive an annual stipend (bursary) of £21,237, including inner London weighting, plus payment of their full-time home tuition fees for a period of 42 months (3.5 years).

You should be eligible for home (UK) tuition fees there are a very limited number (no more than three) of studentships available to overseas applicants, including EU nationals, who meet the academic entry criteria including English Language proficiency.

You will join the internationally recognised researchers in the Department of Computer Science research and PhD programmes | Brunel University London

The Project

Large language models (LLMs) epitomise the current advancements in generative AI, and their astonishing performances outrun a plethora of previously established benchmarks in a variety of tasks. Nonetheless, although LLMs provide a good representation of language generation, they lack thinking and logical capabilities, which often result in faulty reasoning. In an attempt to provide effective solutions, different software tools and prompting strategies have been proposed, but none definitively addressed this shortcoming.

Computational Argumentation is a rich interdisciplinary area of research whose main goal concerns the dialectical formalisation of conflicting data to elaborate an acceptable synthesis of information. Argumentative reasoning is particularly suited for models that parse, work and generate natural language and may offer an intuitive and successful approach to improve LLM reasoning capabilities.

This research project aims to explore, develop and evaluate a reasoning engine driven by computational argumentation capable of enhancing LLMs' thinking and logical faculties.

Please contact Dr Federico Castagna at Federico.Castagna@brunel.ac.uk for an informal discussion about the studentships.

Eligibility

Applicants will have or be expected to receive a first or upper-second class honours degree in an Engineering, Computer Science, Design, Mathematics, Physics or a similar discipline. A Postgraduate Masters degree is not required but may be an advantage.

Skills and Experience

Applicants will be required to demonstrate the following skills:

• Proficiency in Python programming and related Deep Learning libraries (e.g., TensorFlow, PyTorch);
• Experience working with open-source AI models (e.g., HuggingFace models).

The applicant should be highly motivated, able to work independently as well as in a team, collaborate with others and have good communication skills.

How to apply

There are two stages of the application:

1.Applicants must submit the pre-application form via the following link

https://brunel.onlinesurveys.ac.uk/epsrc-dtp-24-25-pre-application-form-brunel-university-lon

by 16.00 on Friday 5th April 2024.

2.If you are shortlisted for the interview, you will be asked to email the following documentation in a single PDF file to cedps-studentships@brunel.ac.uk within 72hrs.

• Your up-to-date CV;
• Your Undergraduate degree certificate(s) and transcript(s) essential;
• Your Postgraduate Masters degree certificate(s) and transcript(s) if applicable;
• Your valid English Language qualification of IELTS 6.5 overall (minimum 6.0 in each section) or equivalent, if applicable;
• Contact details for TWO referees, one of which can be an academic member of staff in the College.

Applicants should therefore ensure that they have all of this information in case they are shortlisted.

Interviews will take place in April/May 2024.