Solving Inverse Problems of Chaotic Systems with Bidirectional Conditional Flow Matching
Peiyan Hu, Jian Zhang, Jiashu Pan, Ruiqi Feng, Tao Zhang, Zhi-Ming Ma, Yuan-Sen Ting, Gongjie Li, Tailin Wu · 2026
Read Solving Inverse Problems of Chaotic Systems with Bidirectional Conditional Flow Matching by Peiyan Hu, Jian Zhang, Jiashu Pan, Ruiqi Feng, Tao Zhang, Zhi-Ming Ma, Yuan-Sen Ting, Gongjie Li, Tailin Wu (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-2606-24824.txt.
Introduction
Peiyan Hu and eight colleagues attack inverse problems for chaotic systems: infer the starting state from the final one. Chaos makes this hard, since small errors explode exponentially and many starts can end alike. Their method, Bidirectional Conditional Flow Matching, learns maps between the distributions of initial and final states, both ways. Working with distributions captures the randomness of chaotic evolution and limits error blow up. A variant adds conservation law constraints for systems that conserve quantities.
Tests cover the classic Lorenz system, a circuit system, and the high dimensional Lorenz 96 model. The method improves five distribution level metrics over baselines. It also runs more than a hundred times faster. On the three body planet scattering problem, the constrained variant respects conservation laws about as well as ground truth. Real globular cluster observations provide the final test. The authors present the results as a scalable route to chaotic inverse problems.
Catalog record
| CREATOR | Peiyan Hu, Jian Zhang, Jiashu Pan, Ruiqi Feng, Tao Zhang, Zhi-Ming Ma, Yuan-Sen Ting, Gongjie Li, Tailin Wu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 103 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2606.24824
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Solving Inverse Problems of Chaotic Systems with Bidirectional Conditional Flow Matching Peiyan Hu, Jian Zhang, Jiashu Pan, Ruiqi Feng, Tao Zhang, Zhi-Ming Ma, Yuan-Sen Ting, Gongjie Li, Tailin Wu Abstract Modeling chaotic systems is crucial yet challenging. Inverse problems in chaotic dynamics, namely inferring initial conditions from final states, remain largely unsolved because of ill-posedness, non-uniqueness, instability, and potentially chaotic time-reverse dynamics. We address this open problem with Bidirectional Conditional Flow Matching (Bi-CFM), which learns bidirectional mappings between distributions of initial and final states to capture the stochasticity of chaotic evolution and mitigate exponential error accumulation over time. Furthermore, for systems with conservation laws, we extend it to Conservation-constrained Bi-CFM (CBi-CFM). Across the classic Lorenz, Circuit, and high-dimensional Lorenz 96 systems, Bi-CFM improves five distribution-level metrics over baselines while achieving a speedup of more than two orders of magnitude. In the three-body planet-planet scattering problem 1 1 1 Planet-planet scattering is the process that planets gravitationally perturb each other, which leads to orbital variations and instability. in planetary dynamics, CBi-CFM better respects conservation laws, with conservation errors comparable to those of the ground truth. Finally, on real observations of globular clusters, collisional million-body systems shaped by ∼ 10 10 \sim 10^{10} years (10 Gyr) of evolution, our method represents an advance in accuracy, establishing a scalable route to solving inverse problems of long-timescale real-world chaotic dynamics. \intern keywords Chaotic system, Inverse problem, Generative model 1 Introduction Chaotic systems are important yet complex nonlinear physical systems, arising in electrical systems Hasler 2005 ; Sprott 2000 , climate systems Tsonis and Elsner 1989 ; Broecker 1995 , astronomy Laskar and Robutel 1993 ; Contopoulos 2002 ; Barrow 1977 , and biology Olsen and Degn 1977 ; Chialvo et al. 1990 . Their intrinsic complexity poses fundamental challenges for modelling and analysis, primarily because small perturbations in the initial state can be amplified exponentially over time, as illustrated in Fig. 1 c , where trajectories initialized with perturbations of order 10^{-6} rapidly diverge. In this work, we focus on the inverse problem of chaotic systems, namely, inferring initial conditions from target final observations. This long-standing challenge has broad implications in science Pecora and Carroll 1990 ; Kantas et al. 2013 ; Ouannas et al. 2016 and engineering Bennett et al. 1996 ; Gallet et al. 2021 , as recovering initial states is essential for understanding the mechanisms and histories of chaotic physical systems Jung et al. 1999 . Despite its significance, this task is challenging, mainly due to three fundamental reasons. First, such inverse problems are ill-posed, meaning that their solutions may be non-unique or unstable Kaipio and Somersalo 2005 . Second, one of the goals of inverse problems is to infer initial states whose evolved final states are close to target final states. However, errors in the initial states amplify during the forward evolution, leading to larger deviations from the target final states. Third, inverse inference may suffer from information loss when certain bodies or components of the system vanish during evolution. For example, a planet may be scattered out of the planetary system ( Rasio and Ford 1996 ; Goldreich et al. 2004 ; Chatterjee et al. 2008 ) , which causes the loss of its information and makes the reconstruction rely solely on the remaining components. For a deterministic and non-chaotic physical system, once the governing equations of motion are known, traditional numerical methods can infer the initial conditions with small numerical errors. The convergence properties of these methods ensure that by starting from the target final state and integrating the system backward in time, one can obtain a high-fidelity estimate of the initial state whose forward evolution closely matches the target LeVeque 1998 ; Lu et al. 2024 ; Binev et al. 2004 . However, in chaotic systems, randomness arises because deterministic nonlinear dynamics amplify infinitesimal perturbations in the initial conditions into unpredictable behaviors. Consequently, traditional numerical approaches fail to accurately recover the initial state even when the equations are known and integrated backward. In particular, when information loss occurs during forward evolution, for example, when certain bodies or components vanish from the system, traditional numerical methods may become entirely inapplicable. Recent progress in deep learning has opened new possibilities for addressing this challenge. Neural networks have shown the capability to model chaotic dynamics with acceleration, accuracy, and efficiency, leading to a growing number of approaches in this area. However, existing studies mainly focus on other tasks of chaotic systems, such as the prediction task Du et al. 2024 ; Hu et al. 2024 ; Alibert et al. 2025 ; Wang and Li 2024 . Several studies have also focused on using deep learning to solve inverse problems, but they have not focused on the more challenging setting of chaotic systems Raissi et al. 2019 ; Yang et al. 2021 ; Bastek and Kochmann 2023 ; Molinaro et al. 2023 ; Huang et al. 2024 . Figure 1 : Schematic illustration of Bi-CFM and CBi-CFM, and the sensitive dependence on initial conditions. a. Bi-CFM framework. Bi-CFM extends Conditional Flow Matching by jointly learning the forward and inverse conditional distribution p(\mathbf{u}_{T}^{\text{gt}}|\mathbf{u}_{0}^{\text{gt}}) (upper panel) and p(\mathbf{u}_{0}^{\text{gt}}|\mathbf{u}_{T}^{\text{gt}}) (lower panel). The noise-free conditioning state is concatenated with the current flow state as model input. b. CBi-CFM framework. CBi-CFM constrains the probability flow within the conservation manifold \mathcal{M} to enforce conservation quantity H . The prior distribution is sampled via random walks on \mathcal{M} , after which the projection correction is applied to mitigate numerical errors. In addition, the velocity is projected onto the tangent space \mathcal{T}_{\mathbf{u}_{t}^{\prime}}\mathcal{M} , ensuring conservation throughout the flow. c. Exponential sensitivity. In the Lorenz system, perturbations of order 10^{-6} are rapidly amplified, causing initially nearby trajectories to diverge across the chaotic attractor. In this work, as shown in Fig. 1 , we propose Bidirectional Conditional Flow Matching (Bi-CFM), an algorithm for the inverse problem of chaotic systems, and its extended version, Conservation-constrained Bi-CFM (CBi-CFM), which enforces conservation laws when conservation quantities are known in the system. We analyze and address the inverse problem of chaotic systems systematically. The method builds upon the probabilistic generative model Conditional Flow Matching (CFM) Lipman et al. 2023 ; Liu et al. 2023 , the advantages of which stem from two key insights. First, under limited computational resources, a chaotic system can be regarded as a stochastic system Berliner 1992 ; Leith 1996 ; Stone and Leigh 2019 , and we therefore learn the distribution of its states rather than the deterministic trajectory. This probabilistic modeling framework is also more suitable for addressing the ill-posed problem, since the solutions may not be unique. Second, we directly learn the end-to-end mapping between initial and final states, rather than relying on time-dependent iterative updates as in traditional numerical solvers. This design avoids repeatedly accumulating numerical errors along the temporal dimension, making the inference error more controllable even as the horizon length increases. Without iterations proportional to the horizon length, this design mitigates the error accumulation in chaotic dynamics. In addition, we introduce a bidirectional modeling framework that jointly learns both forward and reverse dynamics, which allows the model to capture the global structure of chaotic systems rather than focusing solely on reproducing initial states. By enforcing