SURGE : Approximation and Training Free Particle Filter for Diffusion Surrogate
Lifu Wei, Yinuo Ren, Naichen Shi, Yiping Lu · 2026
Read SURGE : Approximation and Training Free Particle Filter for Diffusion Surrogate by Lifu Wei, Yinuo Ren, Naichen Shi, Yiping Lu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-18745.txt.
Introduction
Lifu Wei and three co-authors present SURGE, a 2026 particle filter for diffusion stand-ins. Data assimilation tracks a system state step by step from noisy partial observations. Their method uses a diffusion model as a world model to mimic motion, then guides it with likelihood toward matching states. Since guidance alone cannot guarantee true sampling, Sequential Monte Carlo over the path redraws samples toward the desired target. The result is an unbiased particle filter fusing data with simulations, needing no approximation or training.
Catalog record
| CREATOR | Lifu Wei, Yinuo Ren, Naichen Shi, Yiping Lu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 73202 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.18745
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
SURGE : Approximation and Training Free Particle Filter for Diffusion Surrogate Lifu Wei, Yinuo Ren, Naichen Shi, Yiping Lu Abstract Data assimilation (DA) tackles the sequential estimation of a dynamical system’s latent state from noisy, partial observations. In this paper, we study DA in a setting where the system dynamics are represented by a pretrained diffusion model used as a surrogate forecaster. We focus on how to integrate incoming observations into the diffusion surrogate’s predictions to support continuous state correction and progressively refining the estimated trajectory over time. After receiving noisy observations, the diffusion model is guided using the observation likelihood to steer the generation process toward observation-consistent states. However, such guidance does not guarantee sampling from the true posterior. Motivated by particle filtering methods, we represent the posterior distribution using an ensemble of particles. We perform Sequential Monte Carlo over diffusion trajectories, working with their path measure. We compute importance weights for generated particles and resample to focus on trajectories consistent with the observations. This procedure corrects the generation dynamics, drives the particle approximation toward the desired posterior and leads to an approximation-free particle filtering method that rigorously fuses observational data with diffusion model simulations. \svgpath figures/ 1 Introduction Recovering the state of complex dynamical systems from noisy and incomplete observations is a fundamental problem in science and engineering. Accurate state estimation underpins reliable prediction in applications such as weather forecasting, oceanography, seismology, and fluid dynamics, where system evolution is governed by nonlinear, high-dimensional, and often stochastic partial differential equations. Data assimilation (DA) addresses this challenge by combining model-based forecasts with observational data to produce physically consistent state estimates. Figure 1 : Many modern systems come with complementary “digital” and “physical” twins: a diffusion-based digital twin learns from historical trajectories to generate probabilistic forecasts over dynamics, while sparse and noisy measurements from the physical twin are assimilated to correct and fuse multiple predictions into a single posterior state estimate. In many modern settings, we increasingly have access to two complementary “twins” of the same system ( 43 ) . On one hand, a data-driven digital twin can be learned from historical trajectories, for which diffusion models are a particularly suitable choice because they can represent high-dimensional complex distributions and generate realistic stochastic evolutions ( 26 ; 21 ; 33 ; 41 ) . On the other hand, a physical twin delivers measurements from the real system, often sparse, noisy, and incomplete, but grounded in reality. This motivates a DA perspective in which a diffusion-based digital twin produces probabilistic forecasts, while observations from the physical twin are used to correct and aggregate multiple predictive sources into a single posterior estimate of the system state. However, when the underlying dynamics are simulated by a diffusion model rather than an explicit mechanistic time-stepping solver, it is nontrivial to design a filtering/assimilation framework that incorporates physical observations in a principled way. In particular, a key challenge is how to fuse the diffusion model’s prior forecast distribution with the physical twin’s observations to obtain an approximation-free (or controlled-bias), physically consistent posterior, while remaining computationally tractable in the high-dimensional, nonlinear regime. Mathematically, a discrete-time stochastic dynamical system can be described by: \displaystyle x_{t+1}=\Psi(x_{t},\xi_{t}),\qquad y_{t+1}=\mathcal{A}(x_{t+1})+\varepsilon_{t+1}, (1) wher x_{t}\in\mathbb{R}^{D} is the state vector at time step k , \Psi is the state transition map and \xi_{t} denotes the stochastic force. The observations y_{t}\in\mathbb{R}^{M} are related to the state through the measurement map \mathcal{A} , with observation noise \varepsilon_{t+1} . DA seeks to infer the posterior of the state trajectory x_{0:T} given observations y_{1:T} . This paper focuses on the case where a dynamic system is simulated using a generative digital twin. More specifically, the conditional transition density p(x_{t+1}\mid x_{t}) of dynamical system with transition x_{t+1}=\Psi(x_{t},\xi_{t}) is approximated using a conditional diffusion model ( 52 ; 34 ; 36 ; 2 ; 1 ) . Sampling the next state is implemented by simulating an internal diffusion time process that maps a base distrbution to a sample in state space. When a new observation arrives, a common practice is to guide diffusion sampling by the observation likelihood, swterring samples toward observation-consistent states. However, unless the guidance is the exact Doob’s h -transform ( 15 ; 55 ; 39 ; 45 ; 48 ; 4 ; 16 ; 23 ; 35 ; 66 ) , guided sampling does not in general produce samples from the correct posterior. Exact Doob guidance typically requires solving an associated backward Kolmogorov equation ( 3 ; 65 ) , which is intractable in high-dimensional generative models. In most cases, the resulting guidance is necessarily approximate and often heuristic ( 44 ; 9 ) . A natural question is thus: How can we fuse diffusion-based digital twin simulations with partial observations without introducing systematic bias from approximate guidance? To construct an approximation-free data assimilation framework, we draw inspirations from particle filtering, which approximates the filtering posterior by propagating and reweighting an ensemble of particles according to the system dynamics and the observation likelihood. In our approach, we evaluate the probability of each generated trajectory using the Girsanov change of measure, which accounts for the discrepancy between the guided diffusion model dynamics and the target dynamics. We perform Sequential Monte Carlo (SMC) updates at each step to reweight and resample particles. This sequential resampling mitigates particle degeneracy and ensures that the particle ensemble remains concentrated in regions of high posterior probability. To reduce resampling variance, we incorporate the observation likelihood progressively along the trajectory, so that particle weights remain well-balanced at each small time step rather than becoming highly concentrated due to a single, large likelihood update. Building on this idea, we propose the SURGE filter ( S equential U nbiased R esampling via G irsanov E stimation), a particle filtering-based framework that enables approximation-free data assimilation for diffusion-model-based simulators. Operating exclusively at the inference phase, SURGE is designed to enhance performance in high-dimensional, long-horizon physical processes characterized by sparse and noisy observations. Figure 2 : Conceptual description of SURGE for Data Assimilation. We present SURGE, an approximation-free data assimilation framework. SURGE presents diffusion-based process as a path distribution and perform reweighting and resampling on multi stochastic trajectories during inference to achieve approximation-free output. Contributions. Our contributions are as follows: • We introduce SURGE , an approximation-free data assimilation method that rigorously combines a diffusion surrogate-based digital twin with noisy, partial measurements from a physical twin via particle filtering, enabling sequential state correction over time. • We cast diffusion-based forecasting as inference over a path distribution and perform reweighting and resampling on diffusion trajectories, using a Girsanov change-of-measure to compute importance weights that correct bias introduced by likelihood guidance. We further stabilize filtering in high dimensions via progressive (time-distributed) likelihood incorporation, mitigating particle degeneracy and rigorously fusing observations with diffusion-model simulations. • We apply SURGE solely at the inference level of prevailing diffusion-based data assimilation methods, evaluating it under conditions of long-time sequence prediction and high-dimensional chaotic dynamics. The findings confirm that SURGE adeptly governs sequential generation for diffusion surrogates, enhancing performance during the inference phase. 2 Preliminaries In this section, we will briefly review the background of data assimilation and particle filtering. 2.1 Data Assimilation Given the stochastic dynamical system ( 1 ), the goal of