Search arXivSearch

arXiv · 2405.12453

One-step data-driven generative model via Schrödinger Bridge

Abstract

Generating samples from a probability distribution is a fundamental task in machine learning and statistics. This article proposes a novel scheme for sampling from a distribution for which the probability density $μ({\bf x})$ for ${\bf x}\in{\mathbb{R}}^d$ is unknown, but finite independent samples are given. We focus on constructing a Schrödinger Bridge (SB) diffusion process on finite horizon $t\in[0,1]$ which induces a probability evolution starting from a fixed point at $t=0$ and ending with the desired target distribution $μ({\bf x})$ at $t=1$. The diffusion process is characterized by a stochastic differential equation whose drift function can be solely estimated from data samples through a simple one-step procedure. Compared to the classical iterative schemes developed for the SB problem, the methodology of this article is quite simple, efficient, and computationally inexpensive as it does not require the training of neural network and thus circumvents many of the challenges in building the network architecture. The performance of our new generative model is evaluated through a series of numerical experiments on multi-modal low-dimensional simulated data and high-dimensional benchmark image data. Experimental results indicate that the synthetic samples generated from our SB Bridge based algorithm are comparable with the samples generated from the state-of-the-art methods in the field. Our formulation opens up new opportunities for developing efficient diffusion models that can be directly applied to large scale real-world data.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Hanwen Huang. 2024-05-21. One-step data-driven generative model via Schrödinger Bridge. https://arxiv.org/abs/2405.12453

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

Penguin data reanalyzed via Computational Taxonomy

We employ Computational Taxonomy (CT) to reanalyze the penguin data set penguins_lter by validating and addressing two biological issues: Sexual Size Dimorphism (SSD) and mate-selection criteria. Via Scientific Data Analysis (SDA) computing, CT constructs a Taxonomic Hierarchy by splitting Species first and then Sex, without involving Island, to achieve less complexity. This Taxonomic Hierarchy validates SSD as a branch comparison: (Species, Sex = Male)-vs-(Species, Sex = Female), upon which SDA explores all potential pieces of associative information from all covariate feature-sets, including interacting effects from order-2 to order-4, and then confirms them via their idiosyncratic reliability checks. The collective of confirmed information pieces are displayed on a heatmap platform to manifest underlying dynamics of SSD with explicit block-structured heterogeneity found within males and females. SSD dynamics is explained through mechanistic dependence pertaining to one chief factor consisting of up to 8 feature-sets: Body-Mass coupled by combinations of {Culmen-length,Culmen-depth, Flipper-length}, and two minor factors consisting of low-order combinations of {Culmen-length,Culmen-depth, Flipper-length}. Such Intra-Sex heterogeneity invalidates all Logistic regression modeling on SSD in the original paper. Further, we explore potential mate-selection criteria through the data-frame of Nest-ID within-species homogeneity.

stat.CO

Fast inversion of the generalized Fisher transformation of correlation matrices

The generalized Fisher transformation maps a non-singular correlation matrix to an unconstrained real vector through the off-diagonal elements of its matrix logarithm. Evaluating its inverse is a computational bottleneck in dynamic correlation and multivariate volatility models. We develop a fast inversion algorithm by characterizing the unknown diagonal as the minimizer of a smooth, strictly convex, and coercive objective. An explicit Hessian and global spectral bounds identify the standard fixed-point iteration as a quasi-Newton method and explain why it can converge slowly near singularity. Every fixed-point step decreases the objective, and the iteration converges from every starting point. These results motivate GFT-FP+N, a hybrid of fixed-point and matrix-free Newton steps that never forms the Jacobian. In benchmarks with up to 1,000 replications per design and dimensions up to 800, GFT-FP+N reduces computation time by up to a factor of forty-five relative to the fixed-point iteration and converged in every replication, including on designs where Broyden's method almost always fails. Julia and R packages are provided.

stat.CO

Exact Simulation of Diffusions via Brownian Bridge Range Reconstruction

We develop an exact simulation algorithm for scalar diffusion paths and diffusion bridges when the Poisson potential is unbounded in both tails. The method reconstructs the realized range of a Brownian bridge proposal by sampling its maximum and location, together with the maxima and locations of the two adjacent restricted Brownian meanders. Conditional on this finite information, the remaining path decomposes into four conditionally independent interval-constrained Brownian bridges, which can be sampled exactly at the Poisson times required by the rejection test. In contrast to constructions based on an enclosing range layer, the proposed representation retains the exact extrema and their locations. Our algorithm returns an exact finite-dimensional skeleton without time-discretization error and permits exact post-acceptance refinement at arbitrary finite collections of times. Numerical experiments validate the resulting finite-dimensional laws and identify the restricted-meander extremum simulation as the principal computational cost in the nonlinear example.

stat.CO