Timezone: »

A New Look on Diffusion Times for Score-based Generative Models
Giulio Franzese · Simone Rossi · Lixuan YANG · alessandro finamore · Dario Rossi · Maurizio Filippone · Pietro Michiardi
Score-based diffusion models map noise into data using stochastic differential equations. While current practice advocates for a large $T$ to ensure closeness to steady state, a smaller value of $T$ should be preferred for a better approximation of the score-matching objective and computational efficiency. We conjecture, contrary to current belief and corroborated by numerical evidence, that the optimal diffusion times are smaller than current practice.

#### Author Information

##### Simone Rossi (EURECOM)

PhD Student in Bayesian Deep Learning at Sorbonne Universite/EURECOM (previously MSc in Electronic Engineering and MSc in Computer Engineering). Under the supervision of Prof. Maurizio Filippone, I'm investigating new and exciting problems in Deep Probabilistic Modelling, Approximate Inference, Bayesian Deep Learning and Variational Inference.

##### Dario Rossi (Huawei Technologies France SASU)

Dario Rossi is Director of Huawei AI4NET Lab and Director of the DataCom Lab at the Paris Research Center, France. Before joining Huawei in 2018, he held Full Professor positions at Telecom Paris and Ecole Polytechnique and was holder of Cisco’s Chair NewNet@Paris. He has coauthored 15 patents and over 200 papers in leading conferences and journals, that attracted over 7000 citations and received 9 best paper awards. He is a Senior Member of IEEE and ACM, and receipient of Google Faculty Research Award (2015) and IRTF Applied Network Research Prize (2016)

##### Pietro Michiardi (EURECOM)

Pietro Michiardi received his M.S. in Computer Science from EURECOM and his M.S. in Electrical Engineering from Politecnico di Torino. Pietro received his Ph.D. in Computer Science from Telecom ParisTech (former ENST, Paris), and his HDR (Habilitation) from UNSA. Today, Pietro is a Professor of Computer Science at EURECOM, where he leads the Distributed System Group, which blends theory and system research focusing on large-scale distributed systems (including data processing and data storage), and scalable algorithm design to mine massive amounts of data. Additional research interests are on system, algorithmic, and performance evaluation aspects of distributed systems. Pietro has been appointed as Data Science department head in May 2016.