arXiv subjects

Alexander Scheinker

Publications and source records attributed to Alexander Scheinker.

At least 19 recordsLinked to original sources

arXiv2026

Round-Trip Consistency: Bidirectional Diffusion Models Can Predict Their Own Rollout Errors

Autoregressive models accumulate error over long rollouts, yet at deployment there is no ground truth to measure it against. We train a single conditional latent diffusion model that steps a dynamical system forward or backward in time via a direction flag, and show that this bidirectionality supplies a measurement-free test-time error signal: rolling forward $i$ steps and then backward $i$ steps must return the model to its start, so the round-trip discrepancy $\mathcal{C}_i$ is a self-supervised proxy for the unobservable rollout error: no ensembles, no held-out data, no governing equations, for one extra rollout. We validate on compressible magnetohydrodynamics (MHD), an astrophysical turbulent radiative mixing layer, and natural face videos (CelebV-HQ). On held-out MHD trajectories, $\mathcal{C}_i$ ranks rollout error (Spearman $0.91$-$0.98$ at fixed depth; $0.69 \pm 0.16$ within trajectories), and a simple calibrator fit on training rollouts predicts its magnitude to within $1.14\times$ ($68\%$) and $1.29\times$ ($95\%$) with near-nominal coverage - one nat beyond a depth-only predictor, transferring to all six decoded physical fields. The same signal flags the out-of-distribution Orszag-Tang vortex (AUROC $0.98$; $1.0$ by depth $10$) exactly where sampling-dispersion baselines invert, and it cuts incurred error by $15\%$ at $80\%$ coverage - three times the depth-only baseline. Bidirectional training comes at negative cost, beating direction specialists in both directions, and the backward direction doubles as a fast inverse solver. On LE-PDE-UQ's turbulent Navier-Stokes benchmark, a single bidirectional model reaches accuracy within $1.3\times$ of their ten-model ensemble at a tenth of the training cost, with the best training-free pixel-level calibration. Round-trip consistency turns reversibility into a practical trust signal for generative models.

stat.ML↗

arXiv2026

Bidirectional Autoregressive Latent Diffusion for Forward and Inverse Magnetohydrodynamics

This work presents a new bidirectional autoregressive latent diffusion approach for predicting the evolution of multiple fields (mass density, pressure, velocity, and magnetic field components) for magnetohydrodynamics. We show that this bidirectional flow can be used as a self-supervised consistency metric for uncertainty and error estimation, which enables the model to estimate test-time uncertainty and error without access to ground truth, by comparing how closely flowing forwards and backwards in time returns to the same predicted fields. We also demonstrate this methods's potential to serve as a non-invasive plasma diagnostic, and show how adaptive feedback can be used to make the model more robust based on sparse diagnostics or limited views/measurements.

stat.ML↗

arXiv2026

Mahalanobis-Guided Latent OOD Detection for Hybrid ES-DRL Control in Time-Varying Systems

In this paper, we study Mahalanobis-guided latent out-of-distribution (OOD) detection for test-time RL controller switching in nonlinear time-varying systems. RL controllers can quickly control high-dimensional systems within the training distribution, but their performance can degrade when time-varying dynamics produce unseen observations. We consider a combined ES--DRL controller, where RL provides fast in-distribution actions and bounded extremum seeking (ES) provides robust model-independent control under OOD operation. The key challenge is deciding when to switch. We train a variational autoencoder (VAE) on in-distribution beam-profile observations and use Mahalanobis distance in the VAE latent space to detect OOD beam profiles at test time. This OOD decision sets a binary switch that selects either the RL controller or the ES controller. We evaluate the approach in safety-critical particle accelerator control. In this setting, spatial magnet motion creates OOD beam profiles that were not seen during RL training. Visualization of the VAE latent space shows that the proposed method identifies this OOD scenario and provides an interpretable signal for switching between RL and ES in the combined controller.

cs.LG↗

arXiv2026

PhaseFlow4D: Physically Constrained 4D Beam Reconstruction via Feedback-Guided Latent Diffusion

We address the problem of recovering a time-varying 4D distribution from a sparse sequence of 2D projections - analogous to novel-view synthesis from sparse cameras, but applied to the 4D transverse phase space density $\rho(x,p_x,y,p_y)$ of charged particle beams. Direct single shot measurement of this high-dimensional distribution is physically impossible in real particle accelerator systems; only limited 1D or 2D projections are accessible. We propose PhaseFlow4D, a feedback-guided latent diffusion model that reconstructs and tracks the full 4D phase space from incomplete 2D observations alone, with built-in hard physics constraints. Our core technical contribution is a 4D VAE whose decoder generates the full 4D phase space tensor, from which 2D projections are analytically computed and compared against 2D beam measurements. This projection-consistency constraint guarantees physical correctness by construction - not as a soft penalty, but as an architectural prior. An adaptive feedback loop then continuously tunes the conditioning vector of the latent diffusion model to track time-varying distributions online without retraining. We validate on multi-particle simulations of heavy-ion beams at the Facility for Rare Isotope Beams (FRIB), where full physics simulations require $\sim$6 hours on a 100-core HPC system. PhaseFlow4D achieves accurate 4D reconstructions 11000$\times$ faster while faithfully tracking distribution shifts under time-varying source conditions - demonstrating that principled generative reconstruction under incomplete observations transfers robustly beyond visual domains.

physics.acc-ph↗

arXiv2026

Deep Reinforcement Learning for Robotic Manipulation under Distribution Shift with Bounded Extremum Seeking

Reinforcement learning has shown strong performance in robotic manipulation, but learned policies often degrade in performance when test conditions differ from the training distribution. This limitation is especially important in contact-rich tasks such as pushing and pick-and-place, where changes in goals, contact conditions, or robot dynamics can drive the system out-of-distribution at inference time. In this paper, we investigate a hybrid controller that combines reinforcement learning with bounded extremum seeking to improve robustness under such conditions. In the proposed approach, deep deterministic policy gradient (DDPG) policies are trained under standard conditions on the robotic pushing and pick-and-place tasks, and are then combined with bounded ES during deployment. The RL policy provides fast manipulation behavior, while bounded ES ensures robustness of the overall controller to time variations when operating conditions depart from those seen during training. The resulting controller is evaluated under several out-of-distribution settings, including time-varying goals and spatially varying friction patches.

cs.RO↗

arXiv2026

Nested Extremum Seeking Converges to Stackelberg Equilibrium

The nested Extremum Seeking (nES) algorithm is a model-free optimization method that has been shown to converge to a neighborhood of a Nash equilibrium. In this work, we demonstrate that the same nES dynamics can instead be made to converge to a neighborhood of a Stackelberg (leader--follower) equilibrium by imposing a different scaling law on the algorithm's design parameters. For the two--level nested case, using Lie--bracket averaging and singular perturbation arguments, we provide a rigorous stability proof showing semi-global practical asymptotic convergence to a Stackelberg equilibrium under appropriate time-scale separation. The results reveal that equilibrium selection, Nash versus Stackelberg, depends not on modifying the closed-loop dynamics, but on the hierarchical scaling of design parameters and the induced time-scale structure. We demonstrate this effect using a simple quadratic example and the canonical Fish War game. The Stackelberg variant of nES provides a model-free framework for hierarchical optimization in multi-time-scale systems, with potential applications in power grids, networked dynamical systems, and tuning of particle accelerators.

math.OC↗

arXiv2026

Out-of-distribution transfer of PDE foundation models to material dynamics under extreme loading

Most PDE foundation models are pretrained and fine-tuned on fluid-centric benchmarks. Their utility under extreme-loading material dynamics remains unclear. We benchmark out-of-distribution transfer on two discontinuity-dominated regimes in which shocks, evolving interfaces, and fracture produce highly non-smooth fields: shock-driven multi-material interface dynamics (perturbed layered interface or PLI) and dynamic fracture/failure evolution (FRAC). We formulate the downstream task as terminal-state prediction, i.e., learning a long-horizon map that predicts the final state directly from the first snapshot without intermediate supervision. Using a unified training and evaluation protocol, we evaluate two open-source pretrained PDE foundation models, POSEIDON and MORPH, and compare fine-tuning from pretrained weights against training from scratch across training-set sizes to quantify sample efficiency under distribution shift.

cs.LG↗

arXiv2026

PDE foundation model-accelerated inverse estimation of system parameters in inertial confinement fusion

PDE foundation models are typically pretrained on large, diverse corpora of PDE datasets and can be adapted to new settings with limited task-specific data. However, most downstream evaluations focus on forward problems, such as autoregressive rollout prediction. In this work, we study an inverse problem in inertial confinement fusion (ICF): estimating system parameters (inputs) from multi-modal, snapshot-style observations (outputs). Using the open JAG benchmark, which provides hyperspectral X-ray images and scalar observables per simulation, we finetune the PDE foundation model and train a lightweight task-specific head to jointly reconstruct hyperspectral images and regress system parameters. The fine-tuned model achieves accurate hyperspectral reconstruction (test MSE 1.2e-3) and strong parameter-estimation performance (up to R^2=0.995). Data-scaling experiments (5%-100% of the training set) show consistent improvements in both reconstruction and regression losses as the amount of training data increases, with the largest marginal gains in the low-data regime. Finally, finetuning from pretrained MORPH weights outperforms training the same architecture from scratch, demonstrating that foundation-model initialization improves sample efficiency for data-limited inverse problems in ICF.

cs.LG↗

arXiv2026

Advancing accelerator virtual beam diagnostics through latent evolution modeling: an integrated solution to forward, inverse, tuning, and UQ problems

Virtual beam diagnostics relies on computationally intensive beam dynamics simulations where high-dimensional charged particle beams evolve through the accelerator. We propose Latent Evolution Model (LEM), a hybrid machine learning framework with an autoencoder that projects high-dimensional phase spaces into lower-dimensional representations, coupled with transformers to learn temporal dynamics in the latent space. This approach provides a common foundational framework addressing multiple interconnected challenges in beam diagnostics. For \textit{forward modeling}, a Conditional Variational Autoencoder (CVAE) encodes 15 unique projections of the 6D phase space into a latent representation, while a transformer predicts downstream latent states from upstream inputs. For \textit{inverse problems}, we address two distinct challenges: (a) predicting upstream phase spaces from downstream observations by utilizing the same CVAE architecture with transformers trained on reversed temporal sequences along with aleatoric uncertainty quantification, and (b) estimating RF settings from the latent space of the trained LEM using a dedicated dense neural network that maps latent representations to RF parameters. For \textit{tuning problems}, we leverage the trained LEM and RF estimator within a Bayesian optimization framework to determine optimal RF settings that minimize beam loss. This paper summarizes our recent efforts and demonstrates how this unified approach effectively addresses these traditionally separate challenges.

physics.acc-ph↗

arXiv2025

Generalized Multi-Constraint Extremum Seeking

We generalize the Safe Extremum Seeking algorithm to address the minimization of an unknown objective function subject to multiple unknown inequality and equality constraints, relying on recent results of gradient flow systems. These constraints may represent safety or other critical conditions. The proposed ES algorithm functions as a general nonlinear programming tool, offering practical maintenance of all constraints and semiglobal practical asymptotic stability, utilizing a Lyapunov argument on the penalty function and the set-valued Lie derivative. The efficacy of the algorithm is demonstrated on a 2D problem.

math.OC↗

arXiv2025

Improved Robustness of Deep Reinforcement Learning for Control of Time-Varying Systems by Bounded Extremum Seeking

In this paper, we study the use of robust model independent bounded extremum seeking (ES) feedback control to improve the robustness of deep reinforcement learning (DRL) controllers for a class of nonlinear time-varying systems. DRL has the potential to learn from large datasets to quickly control or optimize the outputs of many-parameter systems, but its performance degrades catastrophically when the system model changes rapidly over time. Bounded ES can handle time-varying systems with unknown control directions, but its convergence speed slows down as the number of tuned parameters increases and, like all local adaptive methods, it can get stuck in local minima. We demonstrate that together, DRL and bounded ES result in a hybrid controller whose performance exceeds the sum of its parts with DRL taking advantage of historical data to learn how to quickly control a many-parameter system to a desired setpoint while bounded ES ensures its robustness to time variations. We present a numerical study of a general time-varying system and a combined ES-DRL controller for automatic tuning of the Low Energy Beam Transport section at the Los Alamos Neutron Science Center linear particle accelerator.

cs.LG↗

arXiv2025

MORPH: PDE Foundation Models with Arbitrary Data Modality

We introduce MORPH, a modality-agnostic, autoregressive foundation model for partial differential equations (PDEs). MORPH is built on a convolutional vision transformer backbone that seamlessly handles heterogeneous spatiotemporal datasets of varying data modality (1D--3D) at different resolutions, and multiple fields with mixed scalar and vector components. The architecture combines (i) component-wise convolution, which jointly processes scalar and vector channels to capture local interactions, (ii) inter-field cross-attention, which models and selectively propagates information between different physical fields, (iii) axial attentions, which factorize full spatiotemporal self-attention along individual spatial and temporal axes to reduce computational burden while retaining expressivity. We pretrain multiple model variants on a diverse collection of heterogeneous PDE datasets and evaluate transfer to a range of downstream prediction tasks. Using both full-model fine-tuning and parameter-efficient low-rank adapters, MORPH outperforms models trained from scratch. Across extensive evaluations, MORPH matches or surpasses strong baselines and recent state-of-the-art models. Collectively, these capabilities present a flexible and powerful backbone for learning from the heterogeneous and multimodal nature of scientific observations, charting a path toward scalable and data-efficient scientific machine learning. The source code, datasets, and models are publicly available at https://github.com/lanl/MORPH.

cs.CV↗

arXiv2025

Using Convolutional Neural Networks to Accelerate 3D Coherent Synchrotron Radiation Computations

Calculating the effects of Coherent Synchrotron Radiation (CSR) is one of the most computationally expensive tasks in accelerator physics. Here, we use convolutional neural networks (CNN's), along with a latent conditional diffusion (LCD) model, trained on physics-based simulations to speed up calculations. Specifically, we produce the 3D CSR wakefields generated by electron bunches in circular orbit in the steady-state condition. Two datasets are used for training and testing the models: wakefields generated by three-dimensional Gaussian electron distributions and wakefields from a sum of up to 25 three-dimensional Gaussian distributions. The CNN's are able to accurately produce the 3D wakefields $\sim 250-1000$ times faster than the numerical calculations, while the LCD has a gain of a factor of $\sim 34$. We also test the extrapolation and out-of-distribution generalization ability of the models. They generalize well on distributions with larger spreads than what they were trained on, but struggle with smaller spreads.

physics.acc-ph↗

arXiv2025

Adaptive conditional latent diffusion maps beam loss to 2D phase space projections

Beam loss (BLM) and beam current monitors (BCM) are ubiquitous at particle accelerator around the world. These simple devices provide non-invasive high level beam measurements, but give no insight into the detailed 6D (x,y,z,px,py,pz) beam phase space distributions or dynamics. We show that generative conditional latent diffusion models can learn intricate patterns to map waveforms of tens of BLMs or BCMs along an accelerator to detailed 2D projections of a charged particle beam's 6D phase space density. This transformational method can be used at any particle accelerator to transform simple non-invasive devices into detailed beam phase space diagnostics. We demonstrate this concept via multi-particle simulations of the high intensity beam in the kilometer-long LANSCE linear proton accelerator.

physics.acc-ph↗

arXiv2025

Physics-Informed Super-Resolution Diffusion for 6D Phase Space Diagnostics

Adaptive physics-informed super-resolution diffusion is developed for non-invasive virtual diagnostics of the 6D phase space density of charged particle beams. An adaptive variational autoencoder (VAE) embeds initial beam condition images and scalar measurements to a low-dimensional latent space from which a 326 pixel 6D tensor representation of the beam's 6D phase space density is generated. Projecting from a 6D tensor generates physically consistent 2D projections. Physics-guided super-resolution diffusion transforms low-resolution images of the 6D density to high resolution 256x256 pixel images. Un-supervised adaptive latent space tuning enables tracking of time-varying beams without knowledge of time-varying initial conditions. The method is demonstrated with experimental data and multi-particle simulations at the HiRES UED. The general approach is applicable to a wide range of complex dynamic systems evolving in high-dimensional phase space. The method is shown to be robust to distribution shift without re-training.

cs.LG↗

arXiv2024

QP Based Constrained Optimization for Reliable PINN Training

Physics-Informed Neural Networks (PINNs) have emerged as a powerful tool for integrating physics-based constraints and data to address forward and inverse problems in machine learning. Despite their potential, the implementation of PINNs are hampered by several challenges, including issues related to convergence, stability, and the design of neural networks and loss functions. In this paper, we introduce a novel training scheme that addresses these challenges by framing the training process as a constrained optimization problem. Utilizing a quadratic program (QP)-based gradient descent law, our approach simplifies the design of loss functions and guarantees convergences to optimal neural network parameters. This methodology enables dynamic balancing, over the course of training, between data-based loss and a partial differential equation (PDE) residual loss, ensuring an acceptable level of accuracy while prioritizing the minimization of PDE-based loss. We demonstrate the formulation of the constrained PINNs approach with noisy data, in the context of solving Laplace's equation in a capacitor with complex geometry. This work not only advances the capabilities of PINNs but also provides a framework for their training.

math.OC↗

arXiv2024

Classifier-pruned Bayesian optimization for particle accelerator tuning: Exploring temporally structured manifold of 6D beam phase space

Complex dynamical systems, such as particle accelerators, often require intricate and time-consuming tuning procedures to achieve optimal performance. In many cases, these procedures must also estimate the optimal system parameters governing the dynamics of a spatiotemporal beam, making the task a high-dimensional optimization problem. To address this, we propose a Classifier-pruned Bayesian Optimization-based Latent space Tuner (CBOL-Tuner), a framework for efficient exploration within a temporally-structured latent manifold of 6D beam phase space. The CBOL-Tuner integrates a conditional variational autoencoder for latent space representation, a long short-term memory network for temporal dynamics, a lightweight neural network for parameter estimation, and a classifier-pruned Bayesian optimizer to adaptively search and filter the latent space for optimal solutions.

cs.LG↗

arXiv2024

Time-inversion of spatiotemporal beam dynamics using uncertainty-aware latent evolution reversal

Charged particle dynamics under the influence of electromagnetic fields is a challenging spatiotemporal problem. Many high performance physics-based simulators for predicting behavior in a charged particle beam are computationally expensive, limiting their utility for solving inverse problems online. The problem of estimating upstream six-dimensional phase space given downstream measurements of charged particles in an accelerator is an inverse problem of growing importance. This paper introduces a reverse Latent Evolution Model (rLEM) designed for temporal inversion of forward beam dynamics. In this two-step self-supervised deep learning framework, we utilize a Conditional Variational Autoencoder (CVAE) to project 6D phase space projections of a charged particle beam into a lower-dimensional latent distribution. Subsequently, we autoregressively learn the inverse temporal dynamics in the latent space using a Long Short-Term Memory (LSTM) network. The coupled CVAE-LSTM framework can predict 6D phase space projections across all upstream accelerating sections based on single or multiple downstream phase space measurements as inputs. The proposed model also captures the aleatoric uncertainty of the high-dimensional input data within the latent space. This uncertainty, which reflects potential uncertain measurements at a given module, is propagated through the LSTM to estimate uncertainty bounds for all upstream predictions, demonstrating the robustness of the LSTM against in-distribution variations in the input data.

physics.acc-ph↗