Statistical inference in generative models using scoring rules

Thesis

Statistical models which allow generating simulations without providing access to the density of the distribution are called simulator models. They are commonly developed by scientists to represent natural phenomena and depend on physically meaningful parameters. Analogously, generative networks produce samples from a probability distribution by transforming draws from a noise (or latent) distribution via a neural network; as for simulator models, the density is unavailable. These two fram...

Expand abstract

Authors: Pacchiardi, L

• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: generative neural networks; scoring rules; likelihood-free inference; summary statistics; generative models; deep learning; probabilistic forecasting; generalised Bayesian inference; simulation-based inference; Bayesian inference
• Subjects: Deep learning (Machine learning); Statistics; Neural networks (Computer science); Inference