Stochastic Interpolants

Follow this topic Ask about this topic

Stochastic interpolants describe continuous paths between probability distributions, often carrying simple noise to complex data. Methods learn the path’s velocity or stochastic drift, using tools such as flow matching, score estimation, and optimized schedules; kernel and geometry-aware variants address settings where standard neural models or grids are limiting. They support generation and inference across images, speech, robotics, weather forecasting, and data assimilation, as well as density estimation and scientific modeling. A central tension is balancing fast, few-step sampling and useful uncertainty estimates against numerical error, which can affect sample quality, likelihoods, and physical consistency.

Latest papers

19 papers4 papers in the last four weeks, against 1 the four weeks before. 0.0% of all new papers.

Bootstrap Stochastic Interpolant Tree for Inference-time Search Under Sparse Reward
Oct 16, 2026Binglin Ji, Anindya Sarkar, Yevgeniy VorobeychikInference-Time Search Stochastic Interpolants
Department of Computer Science & Engineering, Washington University in St. Louis, USA
Learning smooth dynamics via Kinematic Interpolants
Oct 15, 2026Yiheng, Jiang, Youssef Marzouk +2Dynamical Systems Nonlinear System Identification
Roger · Massachusetts Institute of Technology. · Imperial College London and Mila – Québec AI Institute. +1
Xaurora: Generative Weather Forecasting with Denoising Stochastic Interpolants from a Foundation Model Prior
Oct 5, 2026Eliot Walt, Miltiadis Kofinas, Nikolaj Mücke +2Probabilistic Forecasting Stochastic Interpolants
Vrije Universiteit Amsterdam · University of Amsterdam · Delft University of Technology +1
DAWIS: Data Assimilation with Windowed Inverse Sampling via Multitask Interpolants
Oct 2, 2026Erik Wikingsson, Martin Andrae, Tomas Landelius +1Data Assimilation Flow-Based Generative Modeling
Linköping University, Sweden · Swedish Meteorological and Hydrological Institute
Generative models for simulation based filtering: Formulations and Empirical Comparisons
Sep 14, 2026Mohammad Al-Jarrah, Wei Deng, Bamdad Hosseini +1Flow Matching Schrödinger Bridges
Department of Aeronautics & Astronautics, University of Washington, Seattle · Department of Applied Mathematics, University of Washington, Seattle · ML Research, Morgan Stanley
I-SplineFlow: Learning Monotone Spline Stochastic Interpolant Schedulers for Few-Step Generation
Aug 23, 2026Md Sakib Hossain Shovon, Md Rifat Ur Rahman, Md Abtahi Majeed Chowdhury +3Few-Step Diffusion Sampling Diffusion Timestep Scheduling
KAIST · Bangladesh University of Engineering and Technology
Temporal Policy: History-Initialized Action Generation for Robotic Learning from Demonstration
Jul 31, 2026Dylan Miller, Martin JagersandFlow Matching Robot Policy Learning
Department of Computing Science, University of Alberta, Canada
Semi-Supervised Conditional Generative Learning through Stochastic Interpolation and Sufficient Representations
Jul 18, 2026Changyu Liu, Yuling Jiao, Jian HuangRepresentation Learning Stochastic Interpolants
Institute for Mathematics and Artificial Intelligence, Wuhan University, Wuhan, China. · School of Artificial Intelligence, Hubei Key Laboratory of Computational Science, Wuhan University, Wuhan, China. · Departments of Data Science and Artificial Intelligence, and Applied Mathematics, The Hong Kong Polytechnic University, Hong Kong, China.
A Theoretical Analysis of Memory and Overfitting Phenomena in Stochastic Interpolation Models
Jun 7, 2026Yunchen Li, Shaohui Lin, Zhou YuStochastic Interpolants Memorization in Generative Models
East China Normal University
Covariance Shrinkage via Stochastic Interpolation
Jun 5, 2026Mathieu Chalvidal, Florentin Coeurdoux, Eric Vanden-EijndenStochastic Interpolants Covariance Estimation
Capital Fund Management 23 Rue de l’Universit´e, 75007 Paris
Midpoint Generative Models
May 28, 2026Daniil Shlenskii, Nikita Gushchin, Lev Novitskiy +2Flow Matching Stochastic Interpolants
1AXXX, Russia · 2Applied AI Institute, Russia · 3Kandinsky Lab, Russia
Learning Energy-Based Models from Stochastic Interpolants using Spatiotemporal Differences
May 26, 2026Hanlin Yu, RuiKang OuYang, Partha Kaushik +3Contrastive Learning Stochastic Interpolants
University of Helsinki · University of Cambridge · Carnegie Mellon University +1
Disentangling Generation and Regression in Stochastic Interpolants for Controllable Image Restoration
May 20, 2026Yi Liu, Jia Ma, Wengen Li +3Diffusion Models for Image Restoration Stochastic Interpolants
Tongji University · Fudan University
Guided Streaming Stochastic Interpolant Policy
May 11, 2026Puming Jiang, Meiyi Wang, Kelvin Lin +2Robot Policy Learning Robot Policy Adaptation
School of Computing, National University of Singapore
Predictive-Generative Drift Decomposition for Speech Enhancement and Separation
May 7, 2026Julius Richter, Yoshiki Masuyama, Christoph Boeddeker +3Stochastic Interpolants Speech Processing
MERL Cambridge, MA, USA
Uncertainty-quantified Pulse Signal Recovery from Facial Video using Regularized Stochastic Interpolants
Apr 12, 2026Vineet R. Shenoy, Cheng Peng, Rama Chellappa +1Remote Photoplethysmography Uncertainty Quantification
Department of Electrical and Computer Engineering, Johns Hopkins University · School of Data Science, University of Virginia
LDDMM stochastic interpolants: an application to domain uncertainty quantification in hemodynamics
Mar 30, 2026Sarah Katz, Francesco Romor, Jia-Jie Zhu +13D Shape Generation Conditional Generative Modeling
WIAS, Weierstrass Institute for Applied Analysis and Stochastics, Anton-Wilhelm-Amo-Str. 39, 10117 Berlin, Germany. · KTH Royal Institute of Technology, Lindstedtsvägen 25, 114 28 Stockholm, Sweden.
Generative Modeling via Kernelized Stochastic Interpolants
Feb 23, 2026Florentin Coeurdoux, Etienne Lempereur, Nathanaël Cuvelle-Magar +2Stochastic Differential Equations Stochastic Interpolants
Capital Fund Management, Paris, France · Département d’informatique, ENS, Université PSL, Paris, France · Collège de France, Paris, France +2
Boltzmann generators for amorphous particle systems
Dec 18, 2025Louis Grenioux, Leonardo Galliano, Ludovic Berthier +2Boltzmann Sampling Continuous Normalizing Flows
CMAP, CNRS, École polytechnique, Institut Polytechnique de Paris, 91120 Palaiseau, France · Center for Computational Mathematics, Flatiron Institute, New York, NY, USA · Laboratoire de Physique de l’École normale supérieure ENS, Université PSL, CNRS, Sorbonne Université, Université de Paris, 75005 Paris, France +2

Oct 16, 2026J/K move · Enter open · S save

Bootstrap Stochastic Interpolant Tree for Inference-time Search Under Sparse Reward

Binglin Ji, Anindya Sarkar, Yevgeniy Vorobeychik

Department of Computer Science & Engineering, Washington University in St. Louis, USA

Inference-Time Search Stochastic Interpolants

Open paper AskPDF arXiv

Tree-based samplers offer a compelling alternative for scalable inference-time search, bypassing the per-step reward and gradient dependencies that introduce systematic bias into SMC-style frameworks. Despite this advantage, contemporary tree samplers fail to achieve broad exploration. Their reliance on learned, DDPM-like transitions inherently couples neighbor states in a trajectory, thereby restricting child-parent feature divergence in the tree sampler. Consequently, these samplers suffer from restricted coverage of the target marginal distribution, severely hindering the exploration critical for efficient search under sparse reward. To overcome these limitations, we introduce a simple yet effective Bootstrap Stochastic Interpolant Tree (BSIT) sampler that decouples child node generation from explicit parent dependencies. By expanding the degrees of freedom during child node sampling, BSIT encourages diverse trajectory rollouts and unlocks broader path-space diversity-the exact catalyst needed for efficient exploration. Crucially, we show that BSIT's improved exploration accelerates convergence to the target distribution while achieving broader distributional coverage than samplers that rely on DDPM-like transitions for child-node generation, all without incurring additional function evaluations or reward feedback, thereby providing a highly efficient basis for rare target search. Extensive experiments across diverse search settings, including rare-species discovery and target-aligned molecule generation tasks, demonstrate BSIT's superior search efficiency over the baselines.