9 papers · 1 filter
Mirror Learning
Yunpeng Liu, Matthew Niedoba, Oluwanifemi A. Adekanye +4
We investigate imitation learning through the lens of third-person observation and propose a framework for mirror learning: acquiring actionable policies from passive observation.…
Integration Matters: Rollout-Based Training for Constrained Diffusion Models
Xiaoxuan Liang, Saeid Naderiparizi, Berend Zwartsenberg +1
Constrained generative models aim to produce samples that satisfy complex feasibility constraints while remaining faithful to the data distribution. Existing constrained generation…
Filtered Posterior Mean Collections: A Unified Framework for Analytical Models of Diffusion Generalization
Matthew Niedoba, Berend Zwartsenberg, Frank Wood
The neural-network denoising functions which form the backbone of image diffusion models are remarkably consistent in their generalization behaviour across a wide variety of networ…
Improved Constrained Generation by Bridging Pretrained Generative Models
Xiaoxuan Liang, Saeid Naderiparizi, Yunpeng Liu +2
Constrained generative modeling is fundamental to applications such as robotic control and autonomous driving, where models must respect physical laws and safety-critical constrain…
Constrained Generative Modeling with Manually Bridged Diffusion Models
Saeid Naderiparizi, Xiaoxuan Liang, Berend Zwartsenberg +1
In this paper we describe a novel framework for diffusion-based generative modeling on constrained spaces. In particular, we introduce manual bridges, a framework that expands the…
Rolling Ahead Diffusion for Traffic Scene Simulation
Yunpeng Liu, Matthew Niedoba, William Harvey +3
Realistic driving simulation requires that NPCs not only mimic natural driving behaviors but also react to the behavior of other simulated agents. Recent developments in diffusion-…
Towards a Mechanistic Explanation of Diffusion Model Generalization
Matthew Niedoba, Berend Zwartsenberg, Kevin Murphy +1
We propose a simple, training-free mechanism which explains the generalization behaviour of diffusion models. By comparing pre-trained diffusion models to their theoretically optim…
Layerwise Proximal Replay: A Proximal Point Method for Online Continual Learning
Jason Yoo, Yunpeng Liu, Frank Wood +1
In online continual learning, a neural network incrementally learns from a non-i.i.d. data stream. Nearly all online continual learning methods employ experience replay to simultan…
Nearest Neighbour Score Estimators for Diffusion Generative Models
Matthew Niedoba, Dylan Green, Saeid Naderiparizi +9
Score function estimation is the cornerstone of both training and sampling from diffusion generative models. Despite this fact, the most commonly used estimators are either biased…