36 papers
This study introduces the enhanced WT-RDF+ framework, which leverages machine learning-assisted parameter tuning to overcome amplitude accuracy limitations in reconstructing Radial Distribution Functions for amorphous Ge-Se and Ag-Ge-Se systems, thereby outperforming standard ML benchmarks even with limited training data.
This paper develops a thermodynamic framework for asymptotic inference that maps statistical concepts like Shannon information and parameter variance to thermodynamic entropy and state variables, revealing fundamental limits on information gain and unifying ensemble and inferential physics as opposing shadow processes.
This paper introduces Partial Decomposition of Granger Causality (PDGC), a novel framework leveraging Partial Information Decomposition to dissect multivariate spectral Granger causality into unique, redundant, and synergistic components, which was successfully applied to physiological networks to reveal distinct patterns of autonomic dysfunction in patients prone to neurally-mediated syncope.
This paper identifies a novel mechanism called "pseudo-coherence," where non-normal pseudospectral amplification in linearly stable, non-oscillatory stochastic systems drives a sharp transition to intermittent collective temporal organization and irreversible currents without requiring intrinsic oscillators or dynamical instabilities.
This paper introduces Extracted Mode Tracking (EMT), an unsupervised machine learning framework that resolves gravity-capillary wave evolution in vessels with unknown boundary conditions by directly extracting wave modes from spatio-temporal data, thereby enabling quantitative analysis of nonlinear dynamics without requiring prior theoretical modeling.
This paper proposes a physics-informed Bayesian latent diffusion model that integrates probabilistic encoding, diffusion dynamics, and physics constraints to achieve stable and reliable anomaly detection for new physics searches in collider data.
This paper demonstrates that utilizing unconventionally large learning rates to induce transient chaotic dynamics during neural network training creates an optimal balance between exploration and exploitation, thereby accelerating convergence to high accuracy across various architectures and tasks.
This paper presents an adaptive entropy-driven sensor selection policy within a camera-LiDAR particle filter that dynamically switches between modalities to optimize tracking accuracy and continuity for single-vessel surveillance, validated through real-world maritime deployment.
This paper demonstrates that Graph Neural Networks outperform traditional models like TabNet in estimating muon particle momentum for the CMS experiment, highlighting the critical role of node feature dimensionality and the benefits of leveraging the data's inherent graph structure to improve trigger system efficiency.
This paper introduces two interpretable symbolic machine learning methods, the Symbolic Neural Forecaster (SyNF) and the Symbolic Tree Forecaster (SyTF), which successfully learn explicit algebraic equations to forecast chaotic time series with accuracy competitive to deep learning while providing transparent insights into the underlying dynamics.
This study introduces a computationally efficient framework that utilizes non-interacting electron density and Bayesian active learning to achieve highly accurate, zero-shot extrapolative predictions of alloy properties across vast compositional landscapes, significantly reducing the data requirements for discovering refractory high-entropy alloys.
This paper introduces a maximum-entropy framework for continuous-time temporal networks that factorizes interactions into global time processes and static edge probabilities, enabling closed-form analytical solutions and improved generative modeling via non-homogeneous Poisson processes.
This paper presents algorithms for estimating Detector Error Models (DEMs) directly from syndrome data without decoders, applying them to Google's Willow chips to reveal that while DEMs optimized for syndrome likelihood better predict unseen data, those optimized for logical performance serve as superior decoder priors, while also uncovering long-range correlated measurement errors and unmodeled artifacts like radiation events.
This paper introduces a novel methodology for analyzing neutron and x-ray scattering event data that bypasses traditional histogramming and least-squares fitting to achieve significantly higher accuracy and efficiency while reducing systematic errors, despite requiring increased computation time and offering a less intuitive approach.
This paper presents a scalable, two-step framework that propagates chemical kinetics parameter uncertainties onto reduced manifolds to enable efficient, spatially resolved uncertainty quantification in high-fidelity reacting flow simulations, revealing significant variability in trajectory and equilibrium times driven by mixing and low-to-intermediate temperature chemistry.
This paper demonstrates that when the prior given hyperparameters in a Bayesian hierarchical model is a canonical maximum entropy distribution, the resulting dependent marginal prior also possesses a maximum entropy property, but with respect to a constraint on the marginal distribution of a function of the unknown parameters, thereby clarifying the implicit information assumptions of such models.
This paper demonstrates that applying Deep Neural Networks to a 2x2 array of linearly-graded Silicon Photomultipliers significantly improves position resolution and linearity while increasing the number of resolvable pixels by a factor of 5.7 to 12.1 compared to traditional reconstruction methods.
This tutorial presents a computationally efficient, two-step Bayesian framework for quantifying uncertainty in linear shock compression data by deriving posterior distributions for model parameters and propagating them through Rankine-Hugoniot equations to generate multiple consistent Hugoniot curves, offering a more robust and interpretable alternative to traditional least squares and bootstrapping methods.
This paper introduces a novel early warning signal called DE-AC, which estimates the dominant eigenvalue from autocorrelation functions to reliably predict the onset of period-doubling bifurcations in cardiac systems with superior sensitivity and specificity compared to existing indicators.
This paper introduces a method that transforms multivariate time series into symbolic sequences to identify statistically significant higher-order dependencies via a Bayesian approach, modeling these motifs as hyperedges in a hypergraph to reveal meaningful interactions in complex systems like neural and social networks.