Latent Space Refinement for Deep Generative Models
arXiv:2106.00792
Abstract
Deep generative models are becoming widely used across science and industry for a variety of purposes. A common challenge is achieving a precise implicit or explicit representation of the data probability density. Recent proposals have suggested using classifier weights to refine the learned density of deep generative models. We extend this idea to all types of generative models and show how latent space refinement via iterated generative modeling can circumvent topological obstructions and improve precision. This methodology also applies to cases were the target model is non-differentiable and has many internal latent dimensions which must be marginalized over before refinement. We demonstrate our Latent Space Refinement (LaSeR) protocol on a variety of examples, focusing on the combinations of Normalizing Flows and Generative Adversarial Networks.
arXiv comment15 pages, 5 figures, 3 tables
References in corpus (34)
- PyTorch: An Imperative Style, High-Performance Deep Learning Library
- Generative Adversarial Networks: An Overview
- Variational Inference with Normalizing Flows
- Normalizing Flows: An Introduction and Review of Current Methods
- Generative Moment Matching Networks
- CaloGAN: Simulating 3D High Energy Particle Showers in Multi-Layer Electromagnetic Calorimeters with Generative Adversarial Networks
- Anomaly Detection with Density Estimation
- Flow-based generative models for Markov chain Monte Carlo in lattice field theory
- Learning New Physics from a Machine
- Event Generation with Normalizing Flows
- How to GAN LHC Events
- OmniFold: A Method to Simultaneously Unfold All Observables
- Getting High: High Fidelity Simulation of High Granularity Calorimeters with High Speed
- i-flow: High-dimensional Integration and Sampling with Normalizing Flows
- Regressive and generative neural networks for scalar field theory
- Neural Networks for Full Phase-space Reweighting and Parameter Tuning
- Fast and accurate simulation of particle detectors using generative adversarial networks
- Unfolding with Generative Adversarial Networks
- Understanding Event-Generation Networks via Uncertainties
- Decoding Photons: Physics in the Latent Space of a BIB-AE Generative Network
- Controlling Physical Attributes in GAN-Accelerated Simulation of Electromagnetic Calorimeters
- Fast simulation of muons produced at the SHiP experiment using Generative Adversarial Networks
- Deep Learning as a Parton Shower
- Lund jet images from generative and cycle-consistent adversarial networks
- Augmented Neural ODEs
- Image-based model parameter optimization using Model-Assisted Generative Adversarial Networks
- E Pluribus Unum Ex Machina: Learning from Many Collider Events at Once
- GANs for generating EFT models
- Flows for simultaneous manifold learning and density estimation
- Efficient sampling of constrained high-dimensional theoretical spaces with machine learning
- Explainable machine learning of the underlying physics of high-energy particle collisions
- Compressing PDF sets using generative adversarial networks
- Data-driven Estimation of Background Distribution through Neural Autoregressive Flows
- Variational Autoencoders for Jet Simulation