Scholar
Stefano Peluchetti
Google Scholar ID: w3Gi3TEAAAAJ
Research Scientist, Sakana AI
Citations & Impact
Citations
333
H-index
9
i10-index
8
Publications
20
Co-authors
0
Publications
Resume (English only)
Academic Achievements
- Introduced bridge matching (BM), a novel method for constructing a diffusion process that transports samples between two target distributions. This method has been successfully employed in state-of-the-art image generation models such as Stable Diffusion 3 and FLUX.1. Further studied BM and introduced its iterative version, iterated bridge matching (I-BM), which converges to a Schrödinger bridge/entropic optimal transport map. Also introduced coupled bridge matching (BM²), a new approach for learning Schrödinger bridges.
Research Experience
- Currently a research scientist at Sakana AI. Previously, he was a principal research scientist at Cogent Labs and a senior quantitative analyst and data scientist at HSBC.
Education
- Obtained an MSc in Economics and a PhD in Statistics from Bocconi University, under the supervision of Gareth O Roberts.
Background
- Research interests include deep learning and statistics, with a focus on the interplay between neural networks and stochastic processes, measure transport via diffusion processes, probabilistic models, and scalable inference.
Co-authors