Nicholas Matthew Boffi

Assistant Professor@CMU, Carnegie Mellon University·美国·OpenReview

研究方向

optimization · control theory · dynamical systems · deep learning · statistical learning · numerical analysis · statistical physics · computational science

6.0

Joint Distillation for Fast Likelihood Evaluation and Sampling in Flow-based Models

ICLR 2026Poster

3.5

Pre-Generating Multi-Difficulty PDE Data For Few-Shot Neural PDE Solvers

ICLR 2026Rejected

7.5

How to build a consistency model: Learning flow maps via self-distillation

NeurIPS 2025Poster

一作

7.3

BoltzNCE: Learning likelihoods for Boltzmann Generation with Stochastic Interpolants and Noise Contrastive Estimation

NeurIPS 2025Poster

三作

6.3

Shallow diffusion networks provably learn hidden low-dimensional structure

ICLR 2025Poster

一作

合作者 (20)

Michael Samuel Albergo

Nicholas Matthew Boffi

Much Ado About Noising: Dispelling the Myths of Generative Robotic Control

Generalised Flow Maps for Few-Step Generative Modelling on Riemannian Manifolds

Test-time scaling of diffusions with flow maps

Sublinear iterations can suffice even for DDPMs

Joint Distillation for Fast Likelihood Evaluation and Sampling in Flow-based Models

Pre-Generating Multi-Difficulty PDE Data For Few-Shot Neural PDE Solvers

How to build a consistency model: Learning flow maps via self-distillation

BoltzNCE: Learning likelihoods for Boltzmann Generation with Stochastic Interpolants and Noise Contrastive Estimation

Shallow diffusion networks provably learn hidden low-dimensional structure