phase transition analysis

Rare events—such as biomolecular conformational changes, phase transitions, and chemical reactions—are notoriously difficult to sample efficiently in computational simulations. This work addresses this challenge by formulating the estimation of the committor function from transition path theory as a stochastic optimal control problem for the first time. It introduces a feedback control strategy based on the committor gradient to actively steer trajectories through reactive regions and designs a novel nonequilibrium sampling protocol to mitigate metastable trapping caused by intermediate potential wells. The approach integrates transition path theory, deep learning optimization, and off-policy learning, employing both direct backpropagation loss and Value Matching loss. Evaluated on benchmark systems, the method significantly improves the accuracy of committor estimation and yields more precise calculations of reaction rates and equilibrium constants, outperforming existing techniques.

Calibrating computational models of cellular signaling to match time-varying biological responses remains challenging, particularly due to difficulties in reconciling simulated dynamics with empirical reference behaviors and disentangling the influence of receptor trafficking mechanisms on signal transduction efficacy. Method: We propose a visual analytics framework integrating time-series graphs and parallel coordinates to jointly map model parameter spaces and dynamic outputs (e.g., time-resolved signal intensity), enabling interactive parameter sensitivity analysis, behavioral plausibility validation, and exploration of how receptor transport pathways modulate signaling efficiency. Contribution/Results: Evaluated on real-world case studies, the framework significantly enhances interdisciplinary collaboration between modelers and biologists. It advances model calibration accuracy, facilitates hypothesis generation regarding mechanistic underpinnings, and improves result interpretability—demonstrating clear innovation in integrative computational biology and systems pharmacology.

Safety verification of decision-making agents in dynamic environments faces challenges including susceptibility to local optima in high-dimensional scenario spaces and difficulty balancing scenario diversity with criticality. Method: This paper proposes a dual-space guided testing framework that jointly optimizes the scenario parameter space and agent behavioral space. It introduces a novel parameter–behavior closed-loop feedback mechanism, integrating hierarchical representation, dimensionality reduction modeling, multi-dimensional subspace evaluation, behavioral criticality quantification, and adaptive mode switching to dynamically balance local perturbation and global exploration. Results: Experiments on five decision-making agents show that the framework increases critical scenario generation by 56.23% on average. It significantly outperforms state-of-the-art methods under a joint parameter–behavior driving metric, achieving superior scenario diversity, coverage, and verification effectiveness.

In chaotic time series, observational signals are highly coupled with underlying dynamical variability, rendering conventional observation-space change-point detection ineffective for identifying mechanistic transitions. To address this, we propose an interpretable parameter-space paradigm: using simulation-based Bayesian inference with a neural posterior estimator to map observed sequences onto dynamic parameter trajectories, followed by standard change-point detection on these trajectories. Our method integrates neural posterior estimation, simulation-based inference, and off-the-shelf change-point algorithms, and is validated on the Lorenz-63 system. Compared to observation-space baselines, it achieves significant improvements in F1 score, localization accuracy, and false positive rate. We further demonstrate posterior identifiability and calibration, robustness to noise and hyperparameter variation, and a unique balance of high statistical accuracy and physical interpretability.

This work proposes a coverage-oriented, end-to-end scientific machine learning framework to address the challenges of accurately modeling multiphysics systems—such as critical heat flux (CHF)—characterized by strong nonlinearity and significant stochasticity. By integrating uncertainty quantification directly into the learning process, the method employs Bayesian heteroscedastic regression, a quality-driven loss function, and conformal prediction to simultaneously optimize predictive accuracy and uncertainty estimation. Crucially, it eliminates the need for post-hoc calibration, enabling the model to dynamically adapt to varying underlying physical mechanisms. This approach not only maintains high predictive fidelity but also yields uncertainty characterizations that are physically consistent, thereby substantially enhancing the capability to model complex multiphysics systems.

This study addresses the challenges of control design in complex industrial processes characterized by multivariable coupled dynamics by proposing an automated control strategy generation framework that integrates large language models (LLMs) with Bayesian optimization. The approach decomposes control design into structured code generation steps, ensuring physical consistency through execution-based validation and feedback-driven repair. It pioneers the automatic synthesis of decentralized PI controller architectures and their tuning environments directly from dynamic process models. Evaluated on a nonlinear gas preheater benchmark, the generated control schemes—subsequently refined via Bayesian optimization—achieve a 26.5% improvement in closed-loop performance and significantly enhance the transient response of pressure loops, thereby demonstrating the method’s effectiveness and novelty.

Existing dynamical system reconstruction models exhibit limited out-of-distribution generalization, particularly when extrapolating across critical points. This work identifies three fundamental structural deficiencies underlying this limitation and introduces an improved framework based on topological feature disentanglement and hierarchical modeling. For the first time, the study derives a closed-form theoretical bound characterizing the reliable extrapolation range of such models. The proposed approach enables high-accuracy, zero-shot predictions in unseen dynamical regimes—such as regions straddling bifurcation points—without requiring additional training, thereby substantially enhancing out-of-distribution generalization performance.

This work addresses abrupt phenomena in continuous generative models—such as mode locking and semantic collapse during sampling—whose underlying mechanisms remain poorly understood. By modeling the denoising process as a gradient flow on a free energy landscape, the study reveals, for the first time from a differential geometric perspective, that such phase transitions originate from projection caustics on the data support: critical regions where the nearest-point projection ceases to be unique. Building on this insight, the authors propose the Critical Boundary Detector (CBD), which accurately identifies unstable windows along generation trajectories, enabling prediction of mode-commitment moments and targeted intervention in geometrically sensitive regions. The method is validated across toy models, standard diffusion frameworks, and latent text-to-image architectures.

This study addresses the limitations of existing SysML verification approaches, which are often tool-dependent and restricted to performance properties, lacking support for automated validation of behavioral and interface requirements. To overcome these shortcomings, this work proposes a tool-agnostic, automated verification workflow driven by SysML test cases, integrating UML Testing Profile and behavioral diagram constructs to enable unified validation of multidimensional attributes—including behavior, timing, and state responses. The methodology was developed through a mixed-methods research strategy combining literature review and stakeholder interviews, and its efficacy was empirically validated across two independent SysML toolchains. The approach not only transcends the constraints of conventional parametric methods but also enables automatic traceability of verification results back to the original model elements.