• Corpus ID: 246634114

Riemannian Score-Based Generative Modeling

@article{Bortoli2022RiemannianSG,
  title={Riemannian Score-Based Generative Modeling},
  author={Valentin De Bortoli and Emile Mathieu and Michael Hutchinson and James Thornton and Yee Whye Teh and A. Doucet},
  journal={ArXiv},
  year={2022},
  volume={abs/2202.02763}
}
Score-based generative models (SGMs) are a powerful class of generative models that exhibit remarkable empirical performance. Score-based generative modelling (SGM) consists of a “noising” stage, whereby a diffusion is used to gradually add Gaussian noise to data, and a generative model, which entails a “denoising” process defined by approximating the time-reversal of the diffusion. Existing SGMs assume that data is supported on a Euclidean space, i.e. a manifold with flat geometry. In many… 

DiffDock: Diffusion Steps, Twists, and Turns for Molecular Docking

Predicting the binding structure of a small molecule ligand to a protein—a task known as molecular docking —is critical to drug design. Recent deep learning methods that treat docking as a regression

From Denoising Diffusions to Denoising Markov Models

A general framework is proposed which not only unifies and generalizes this approach to a wide class of spaces but also leads to an original extension of score matching.

A Survey on Generative Diffusion Model

A diverse range of advanced techniques to speed up the diffusion models – training schedule, training-free sampling, mixed-modeling, and score & diffusion unification are presented.

First Hitting Diffusion Models for Generating Manifold, Graph and Categorical Data

FHDM enables instance-dependent terminate time and accelerates the diffusion process to sample higher quality data with fewer diffusion steps, observing considerable improvement compared with the state-of-the-art approaches in both quality and speed.

Image Restoration with Mean-Reverting Stochastic Differential Equations

This paper presents a stochastic differential equation (SDE) approach for general-purpose image restoration that transforms a high-quality image into a degraded counterpart as a mean state with Gaussian noise and proposes a maximum likelihood objective to learn an optimal reverse trajectory which stabilizes the training and improves the restoration results.

Modeling halo and central galaxy orientations on the SO(3) manifold with score-based generative models

Upcoming cosmological weak lensing surveys are expected to constrain cosmological parameters with unprecedented precision. In preparation for these surveys, large simulations with realistic galaxy

Broadly applicable and accurate protein design by integrating structure prediction networks and diffusion generative models

By fine tuning the RoseTTAFold structure prediction network on protein structure denoising tasks, this work obtains a generative model of protein backbones that achieves outstanding performance on unconditional and topology-constrained protein monomers design, protein binder design, symmetric oligomer design, enzyme active site scaffolding, and symmetric motif scaffolding for therapeutic and metal-binding protein design.

Denoising Deep Generative Models

This paper proposes two methodologies based on adding Gaussian noise to the data to remove the dimensionality mismatch during training, and both provide a denoising mechanism whose goal is to sample from the model as though no noise had been added to theData.

Diffusion Generative Models in Infinite Dimensions

This work generalizes diffusion models to operate directly in function space by developing the foundational theory for such models in terms of Gaussian measures on Hilbert spaces and developing methods for diffusion generative modeling in Sobolev spaces.

The Union of Manifolds Hypothesis

It is argued that the manifold hypothesis does not properly capture the low-dimensional structure typically present in image data, and is empirically verified on commonly-used image datasets, finding that indeed, observed data lies on a disconnected set and that intrinsic dimension is not constant.

References

SHOWING 1-10 OF 151 REFERENCES

TIME REVERSAL OF DIFFUSIONS

On montre que si un processus de diffusion {X t :0≤t≤1} sur R d satisfait dX t =b(t,X t ) dt+σ(t,X t ) dω t alors le processus retourne, {X t :0≤t≤1} ou X t =X 1−t est encore une diffusion avec une

Moser Flow: Divergence-based Generative Modeling on Manifolds

Moser Flow is introduced, a new class of generative models within the family of continuous normalizing flows (CNF) that constitutes a universal density approximator under suitable assumptions and achieves improvements in density estimation, sample quality, and training complexity over existing CNFs on challenging synthetic geometries and real-world benchmarks from the earth and climate sciences.

Time reversal of diffusion processes under a finite entropy condition

. Motivated by entropic optimal transport, time reversal of diffusion processes is revisited. An integration by parts formula is derived for the carré du champ of a Markov process in an abstract

Score-Based Generative Modeling through Stochastic Differential Equations

This work presents a stochastic differential equation (SDE) that smoothly transforms a complex data distribution to a known prior distribution by slowly injecting noise, and a corresponding reverse-time SDE that transforms the prior distribution back into the data distribution by Slowly removing the noise.

Zeitschrift fur Wahrscheinlichkeitstheorie und Verwandte Gebiete.

The central limit problem for geodesic random walks

The purpose of the present work is to consider the problem of defming the concept of a random walk in a general Riemannian manifold JCl, and to investigate the behavior in the limit of a sequence of

Eigenvalues in Riemannian geometry

Introduction to Riemannian Manifolds

Let M (resp. N) be a connected. smooth (= C x ) n-dimensional manifold without boundary. We denote by C x (M) the ring of smooth real valued functions on M and by x(M) the Lie-algebra of all smooth

Stochastic Analysis on Manifolds

The purpose of this chapter is to describe and investigate the main features of stochastic analysis on smooth manifolds. Our principal focus shall be on stochastic differential equations. A
...