13 papers
Controllable Image Generation with Composed Parallel Token Prediction
Jamie Stirling, Noura Al-Moubayed, Chris G. Willcocks +1
Conditional discrete generative models struggle to faithfully compose multiple input conditions. To address this, we derive a theoretically-grounded formulation for composing discr…
Controllable Image Generation with Composed Parallel Token Prediction
Jamie Stirling, Noura Al-Moubayed, Chris G. Willcocks +1
Conditional discrete generative models struggle to faithfully compose multiple input conditions. To address this, we derive a theoretically-grounded formulation for composing discr…
Investigating Permutation-Invariant Discrete Representation Learning for Spatially Aligned Images
Jamie S. J. Stirling, Noura Al-Moubayed, Hubert P. H. Shum
Vector quantization approaches (VQ-VAE, VQ-GAN) learn discrete neural representations of images, but these representations are inherently position-dependent: codes are spatially ar…
Adversarial Defence without Adversarial Defence: Enhancing Language Model Robustness via Instance-level Principal Component Removal
Yang Wang, Chenghao Xiao, Yizhi Li +3
Pre-trained language models (PLMs) have driven substantial progress in natural language processing but remain vulnerable to adversarial attacks, raising concerns about their robust…
AttenCraft: Attention-guided Disentanglement of Multiple Concepts for Text-to-Image Customization
Junjie Shentu, Matthew Watson, Noura Al Moubayed
Text-to-image (T2I) customization empowers users to adapt the T2I diffusion model to new concepts absent in the pre-training dataset. On this basis, capturing multiple new concepts…
Early Detection and Reduction of Memorisation for Domain Adaptation and Instruction Tuning
Dean L. Slack, Noura Al Moubayed
Although large language models excel across many tasks, they can memorise training data and thereby expose private or copyrighted text. Most defences target the pre-training stage,…