This is a project which is currently making use of HPC facilities at Newcastle University. It is active.
For further information about this project, please contact:
This project develops machine learning methods that make large language models and AI agents more reliable, trustworthy and fair. The work covers fine-tuning and evaluating language models, multi-agent systems, and generative models, with a focus on uncertainty quantification, fairness, and domain generalisation.
The project is built mainly in Python using PyTorch and the Hugging Face ecosystem (Transformers, Datasets, etc.). Typical workloads include fine-tuning and inference of large language models, diffusion-based generative models, and agent simulations, all of which benefit from GPU acceleration. Experiment tracking is handled with Weights & Biases, and training uses standard CUDA-based and multi-GPU distributed setups. Jobs range from single-GPU development runs to larger multi-GPU training.