Simone Rossi
Institut Eurécom
14 Papers
13 Citations
Simone Rossi is an academic researcher from Institut Eurécom. The author has contributed to research in topics: Bayesian probability & Inference. The author has an hindex of 4, co-authored 11 publications.
Chat about Author
Papers
•Posted Content
All You Need is a Good Functional Prior for Bayesian Deep Learning
TL;DR: This work proposes a novel and robust framework to match their prior with the functional prior of neural networks based on the minimization of their Wasserstein distance, and provides vast experimental evidence that coupling these priors with scalable Markov chain Monte Carlo sampling offers systematically large performance improvements over alternative choices of priors and state-of-the-art approximate Bayesian deep learning approaches.
49
How Much Is Enough? A Study on Diffusion Times in Score-Based Generative Models
Giulio Franzese,Simone Rossi,Lixuan Yang,Alessandro Finamore,Dario Rossi,Maurizio Filippone,Pietro Michiardi +6 more
TL;DR: This work shows how an auxiliary model can be used to bridge the gap between the ideal and the simulated forward dynamics, followed by a standard reverse diffusion process, and suggests a new method to improve quality and efficiency of both training and sampling, by adopting smaller diffusion times.
•Proceedings Article
Good Initializations of Variational Bayes for Deep Models
Simone Rossi,Pietro Michiardi,Maurizio Filippone +2 more
- 24 May 2019
TL;DR: The authors proposed a layer-wise initialization strategy based on Bayesian linear models for stochastic variational inference, which showed faster and better convergence compared to alternatives inspired by the literature on initializations for loss minimization.
•Posted Content
Sparse Gaussian Processes Revisited: Bayesian Approaches to Inducing-Variable Approximations
TL;DR: This work shows that, by revisiting old model approximations such as the fully-independent training conditionals endowed with powerful sampling-based inference methods, treating both inducing locations and GP hyper-parameters in a Bayesian way can improve performance significantly.
19
Continuous-Time Functional Diffusion Processes
Giulio Franzese,Simone Rossi,Dario Rossi,Markus Heinonen,Maurizio Filippone,Pietro Michiardi +5 more
TL;DR: Functional diffusion processes (FDPs) as mentioned in this paper generalize score-based diffusion models to infinite-dimensional function spaces, which can be used to build a new breed of generative models in function spaces.