Parameter Space Research Articles

Optimizing data acquisition: a Bayesian approach for efficient machine learning model training

Acquiring a substantial number of data points for training accurate machine learning (ML) models is a big challenge in scientific fields where data collection is resource-intensive. Here, we propose a novel approach for constructing a minimal yet highly informative database for training ML models in complex multi-dimensional parameter spaces. To achieve this, we mimic the underlying relation between the output and input parameters using Gaussian process regression (GPR). Using a set of known data, GPR provides predictive means and standard deviation for the unknown data. Given the predicted standard deviation by GPR, we select data points using Bayesian optimization to obtain an efficient database for training ML models. We compare the performance of ML models trained on databases obtained through this method, with databases obtained using traditional approaches. Our results demonstrate that the ML models trained on the database obtained using Bayesian optimization approach consistently outperform the other two databases, achieving high accuracy with a significantly smaller number of data points. Our work contributes to the resource-efficient collection of data in high-dimensional complex parameter spaces, to achieve high precision ML predictions.

Integrating Machine Learning and the Finite Element Method for Assessing Stiffness Degradation in Photovoltaic Modules.

This study introduces a novel machine learning (ML) method utilizing a stacked auto-encoder network to predict stiffness degradation in photovoltaic (PV) modules with pre-existing cracks. The input data for the training process was derived from numerical simulations, ensuring a comprehensive representation of module behavior under various conditions. The findings highlight the robust predictive capability of the model, as evidenced by its impressive R2 value of 0.961 and notably low root mean square error (RMSE) of 4.02%. These metrics significantly outperform those of other conventional methods, including the convolutional neural network (CNN) with R2 of 0.905 and RMSE of 9.43%, the space vector machine (SVM) with R2 of 0.827 and RMSE of 17.93%, and the random forest (RF) with R2 of 0.899 and RMSE of 11.02%. Moreover, the findings suggest that the predictive dynamics of degradation are affected by the varying weight functions of different input parameters, such as climate temperature, grain size, material effort, and pre-crack size, as the degradation level changes. Furthermore, a geometric analysis reveals model deficiencies where significant overestimations correlate with thicker glass components, while pronounced underestimations are predominantly associated with thinner layers of polycrystalline silicon wafer and Ethylene Vinyl Acetate in the module. As a case study, it demonstrated that to maintain a constant degradation level between 1.30 and 1.32 in a PV module with components featuring consistent geometric attributes, the input parameters must be kept within specific ranges: climate temperature ranging from 33 to 57°C, grain size ranging from 36 to 81 μm, material effort ranging from 0.74 to 0.81, and pre-crack size ranging from 24 to 32 μm. Therefore, this underscores that the ML model not only predicts degradation but also delineates the parameter space required to achieve a consistent output value.

Electric Penrose, circular orbits and collisions of charged particles near charged black holes in Kalb–Ramond gravity

General relativity (GR) is a well-tested theory of gravity in strong and weak field regimes. Many modifications to this theory were obtained, including different scalar, vector, and tensor fields to the GR with non-minimal coupling to gravity. Kalb–Ramond (KR) gravity is also a modified theory formulated in the presence of a bosonic field. One astrophysical way to test gravity is by studying the motion of test particles in the spacetime of black holes (BH). In this work, we study the circular motion of charged particles and explore energetic processes around charged BHs in KR theory. First, we investigated the event horizon radius and analyzed horizon-no horizon regions in the BH charge and KR parameter space. Considering the Coulomb interaction, we derive and analyze the effective potential for charged particles around a charged KR BH. We investigate charged particles’ angular momentum and energy corresponding to circular orbits. We also investigate how the KR non-minimal coupling parameter affects the radius of the innermost stable circular orbits, the corresponding energy, and the angular momentum. We also investigated the electric Penrose process and charged-particle collisions near the KR BH. The presence of the nonzero KR parameter results in a decrease in the energy efficiency of the Penrose process. Also obtained is that the KR parameter’s positive (negative) values cause a decrease (increase) in the center of mass energy of colliding particles near the BH horizon.

Measurements of laser-plasma instabilities in double-cone ignition experiments relevant to the direct-drive conditions at Shenguang-II Upgrade laser facility

Experiments have been performed to investigate laser-plasma instabilities (LPIs) relevant to the direct-drive inertial confinement fusion (ICF) conditions at the Shenguang-II Upgrade (SG-II UP) laser facility during the double-cone ignition (DCI) experimental campaigns, with the overlapped laser intensity ranging from 6×1014Wcm−2 to 1.8×1015Wcm−2 . An overall LPI scenario has been built from the collection and investigation of angularly distributed scattered light. Across broad ranges of laser and plasma parameters, relevant to the direct-drive ICF conditions, the dominance of the stimulated Raman side scattering (SRSS) was confirmed, and its coexistence with the two-plasmon decay (TPD) was also observed. Significant SRSS scattered light was observed across an extremely wide range of emission angles, concentrated at large angles, and demonstrated to be robust throughout the parameter space. Time-resolved spectral measurements show distinctly different SRSS behaviors along different angles and a slightly higher threshold compared to the TPD. The overall SRSS energy loss was measured, indicating a scattered light energy fraction of up to 6%. These results are of crucial importance for the precise assessment of the LPIs in the DCI as well as other laser directly driven ICF schemes.

Sensitivity Analysis and Uncertainty of a Myocardial Infarction Model

There is uncertainty in the results of any mathematical model due to different reasons. It is important to estimate this uncertainty. Sensitivity analysis is commonly used to estimate how the changes in the input parameters affect the solutions of the model. In this paper, we discuss different ways of performing local and global sensitivity analyses and apply them to two models: an epidemic model and a new myocardial infarction model, both based on ordinary differential equations. The first model is a simple model used to explain the ideas, while the second one shows how to apply them to a model with more state variables and parameters. We find that if the parameters are not accurately known, local sensitivity analysis can be misleading and that global sensitivity methods that sample the whole parameter space, varying all the values of the parameters at the same time, are the most reliable. We also show how the sensitivity analysis results can be used to determine the uncertainty in the results of the model. We present numerical simulations.

Testing Times: Disentangling Admixture Histories in Recent and Complex Demographies using ancient DNA.

Our knowledge of human evolutionary history has been greatly advanced by paleogenomics. Since the 2020s, the study of ancient DNA has increasingly focused on reconstructing the recent past. However, the accuracy of paleogenomic methods in resolving questions of historical and archaeological importance amidst the increased demographic complexity and decreased genetic differentiation remains an open question. We evaluated the performance and behavior of two commonly used methods, qpAdm and the f3-statistic, on admixture inference under a diversity of demographic models and data conditions. We performed two complementary simulation approaches - firstly exploring a wide demographic parameter space under four simple demographic models of varying complexities and configurations using branch-length data from two chromosomes - and secondly, we analyzed a model of Eurasian history composed of 59 populations using whole-genome data modified with ancient DNA conditions such as SNP ascertainment, data missingness, and pseudo-haploidization. We observe population differentiation is the primary factor driving qpAdm performance. Notably, whilst complex gene-flow histories influence which models are classified as plausible, they do not reduce overall performance. Under conditions reflective of the historical period, qpAdm most frequently identifies the true model as plausible amongst a small candidate set of closely related populations. To increase the utility for resolving fine-scaled hypotheses, we provide a heuristic for further distinguishing between candidate models that incorporates qpAdm model P-values and f3-statistics. Finally, we demonstrate a significant performance increase for qpAdm using whole-genome branch-length f2-statistics, highlighting the potential for improved demographic inference that could be achieved with future advancements in f-statistic estimations.

Invariant amplitudes, unpolarized cross sections, and polarization asymmetries in neutrino-nucleon and antineutrino-nucleon elastic scattering

At leading order in weak and electromagnetic couplings, cross sections for (anti)neutrino-nucleon elastic scattering are determined by four nucleon form factors that depend on the momentum transfer Q2. Including radiative corrections in the Standard Model and potential new physics contributions beyond the Standard Model, eight invariant amplitudes are possible, depending on both Q2 and the (anti)neutrino energy Eν. We review the definition of these amplitudes and use them to compute both unpolarized and polarized observables including radiative corrections. We show that unpolarized accelerator neutrino cross-section measurements can probe new physics parameter space within the constraints inferred from precision beta decay measurements. Published by the American Physical Society 2024

SigAlign: an alignment algorithm guided by explicit similarity criteria.

In biological sequence alignment, prevailing heuristic aligners achieve high-throughput by several approximation techniques, but at the cost of sacrificing the clarity of output criteria and creating complex parameter spaces. To surmount these challenges, we introduce 'SigAlign', a novel alignment algorithm that employs two explicit cutoffs for the results: minimum length and maximum penalty per length, alongside three affine gap penalties. Comparative analyses of SigAlign against leading database search tools (BLASTn, MMseqs2) and read mappers (BWA-MEM, bowtie2, HISAT2, minimap2) highlight its performance in read mapping and database searches. Our research demonstrates that SigAlign not only provides high sensitivity with a non-heuristic approach, but also surpasses the throughput of existing heuristic aligners, particularly for high-accuracy reads or genomes with few repetitive regions. As an open-source library, SigAlign is poised to become a foundational component to provide a transparent and customizable alignment process to new analytical algorithms, toolsand pipelines in bioinformatics.

Blowing in the dark matter wind

Interactions between dark matter and ordinary matter will transfer momentum, and therefore give rise to a force on ordinary matter due to the dark matter ‘wind.’ We show that this force can be maximal in a realistic model of dark matter, meaning that an order-1 fraction of the dark matter momentum incident on a target of ordinary matter is reflected. The model consists of light (mϕ ≲ eV) scalar dark matter with an effective interaction ϕ2ψ¯ψ\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$ {\\phi}^2\\overline{\\psi}\\psi $$\\end{document}, where ψ is an electron or nucleon field. If the coupling is repulsive and sufficiently strong, the field ϕ is excluded from ordinary matter, analogous to the Meissner effect for photons in a superconductor. We show that there is a large region of parameter space that is compatible with existing constraints, where the force is large enough to be detected by existing force probes, such as satellite tests of the equivalence principle and torsion balance experiments. However, shielding of the dark matter by ordinary matter prevents existing experiments from being sensitive to the dark matter force. We show that precise measurements of spacecraft trajectories proposed to test long distance modifications of gravity are sensitive to this force for a wide range of parameters.

Direct production of light scalars in the type-I two-Higgs-doublet model at the lifetime frontier of the LHC

A light pseudoscalar A in the sufficient large tanβ region of type-I two-Higgs-doublet model can be naturally a long-lived particle. We focus on H±A, HA, and AA pair productions via the electroweak processes mediated by the bosons at the LHC, including pp→W±/Z→H±/HA and pp→h→AA at the 14 TeV LHC. The possibility of probing A as a LLP at the FASER-2, FACET, MoEDAL-MAPP-2, and MATHUSLA is discussed. We find that FASER-2 fails to probe any parameter space within 0.2 GeV<mA<10 GeV for all the considered processes. For 130<mH±=mH<400 GeV, FACET, MoEDAL-MAPP-2, and MATHUSLA can probe tanβ≲104−6 for mA≲3 GeV, and tanβ≲106−8 for 3 GeV≲mA<10 GeV from pp→W±/Z→H±/ZA processes. And pp→h→AA process covers similar parameter space. All processes can surpass the current limits. Published by the American Physical Society 2024

Parameter Tuning of Agent-Based Models: Metaheuristic Algorithms

When it comes to modelling complex systems using an agent-based approach, there is a problem of choosing the appropriate parameter optimisation technique. This problem is further aggravated by the fact that the parameter space in complex agent-based systems can have a large dimension, and the time required to perform numerical experiments can be large. An alternative approach to traditional optimisation methods are the so-called metaheuristic algorithms, which provide an approximate solution in an acceptable time. The purpose of this study is to compare various metaheuristic algorithms for parameter tuning and to analyse their effectiveness applied to two agent-based models with different complexities. In this study, we considered commonly used metaheuristic algorithms for agent-based model optimisation: the Markov chain Monte Carlo method, the surrogate modelling approach, the particle swarm optimisation algorithm, and the genetic algorithm, as well as the more novel chaos game optimisation algorithm. The proposed algorithms were tested on two agent-based models, one of which was a simple toy model of the spread of contagious disease, and the other was a more complex model of the circulation of respiratory viruses in a city with 10 million agents and 26 calibrated parameters.

New Physics Opportunities at the DUNE Near Detector

Focusing on elastic neutrino–electron scattering events, we explore the prospect of constraining new physics beyond the Standard Model at the DUNE Near Detector (ND). Specifically, we extract the attainable sensitivities for motivated scenarios such as neutrino generalized interactions (NGIs), the sterile neutrino dipole portal and unitarity violation. We furthermore examine the impact of the τ-optimized flux at the DUNE-ND and compare our results with those obtained using the standard CP-optimized flux. We find that our present analysis is probing a previously unexplored region of the parameter space, complementing existing results from cosmological observations and terrestrial experiments.

Bayesian Federated Learning with Hamiltonian Monte Carlo: Algorithm and Theory

This work introduces a novel and efficient Bayesian federated learning algorithm, namely, the Federated Averaging stochastic Hamiltonian Monte Carlo (FA-HMC), for parameter estimation and uncertainty quantification. We establish rigorous convergence guarantees of FA-HMC on non-iid distributed data sets, under the strong convexity and Hessian smoothness assumptions. Our analysis investigates the effects of parameter space dimension, noise on gradients and momentum, and the frequency of communication (between the central node and local nodes) on the convergence and communication costs of FA-HMC. Beyond that, we establish the tightness of our analysis by showing that the convergence rate cannot be improved even for continuous FA-HMC process. Moreover, extensive empirical studies demonstrate that FA-HMC outperforms the existing Federated Averaging-Langevin Monte Carlo (FA-LD) algorithm.

SECRET: Statistical Emulation for Computational Reverse Engineering and Translation with applications in healthcare

There have been impressive advances in the physical and mathematical modelling of complex physiological systems in the last few decades, with the potential to revolutionise personalised healthcare with patient-specific evidence-based diagnosis, risk assessment and treatment decision support using digital twins. However, practical progress and genuine clinical impact hinge on successful model calibration, parameter estimation and uncertainty quantification, which calls for novel innovative adaptions and methodological extensions of contemporary state-of-the-art inference techniques from Statistics and Machine Learning. In the present study, we focus on two computational fluid-dynamics (CFD) models of the blood systemic and pulmonary circulation. We discuss state-of-the-art emulation techniques based on deep learning and Gaussian processes, which are coupled with established inference techniques based on greedy optimisation, simulated annealing, Markov Chain Monte Carlo, History Matching and rejection sampling for computationally fast inference of unknown parameters of the CFD models from blood flow and pressure data. The inference task was set as a competitive challenge which the participants had to conduct within a limited time frame representative of clinical requirements. The performance of the methods was assessed independently and objectively by the challenge organisers, based on a ground truth that was unknown to the method developers. Our results indicate that for the systemic challenge, in which an idealised case of noise-free data was considered, the relative deviation from the ground-truth in parameter space ranges from 10−5% (highest-performing method) to 3% (lowest-performing method). For the pulmonary challenge, for which noisy data was generated, the performance ranges from 0.9% to 7% deviation for the parameter posterior mean, and from 35% to 570% deviation for the parameter posterior variance.

Causality criteria from stability analysis at ultrahigh boost

In this work, we have exclusively employed the linear stability analysis at ultrahigh boost on two well-known stable-causal theories—second-order Müller-Israel-Stewart and first-order Bemfica-Disconzi-Noronha-Kovtun, to identify the region of parameter space over which they are frame-invariantly stable and obey causal signal propagation. It has been shown for these two theories that at near-luminal boost, stability criteria alone can provide the causality constraints on transport coefficients, which are identical to the asymptotic causality conditions, without actually going to the asymptotic limit of the theories. The analysis indicates an alternative approach to derive the causality constraints, which is more appropriate for low-energy effective theories like relativistic hydrodynamics. Published by the American Physical Society 2024

Tunable gelatin methacrylate polyethylene glycol diacrylate hydrogels for cell mechanosensing applications

Three-dimensional (3D) tissue-engineered scaffolds mimic the physiological environment of cells by providing essential structural support, biochemical cues, and the mechanical strength needed for cell adhesion, proliferation, migration, and differentiation. Hydrogels like polyethylene glycol diacrylate (PEGDA) are commonly used biomaterials for cell culture due to their affordability, tunable stiffness, and ability to efficiently transport nutrients and gases. However, PEGDA lacks cell adhesion sites essential for cell proliferation and migration and has limited degradability. Methacrylated gelatin (GelMA) produced from denatured bovine collagen, crosslinks under ultraviolet light (UV) resulting in a degradable hydrogel with cell adhesion sites. Here, we synthesized GelMA with variable degree of methacrylation and crosslinked it with PEGDA to produce cell scaffolds with independently tunable mechanical and biochemical properties by varying the ratios of the two polymers. We determined polymer ratios that resulted in scaffolds with different mechanical properties but the same gelatin concentrations (providing cell adhesion and degradation sites) as well as different gelatin concentrations but the same mechanical properties. With the developed scaffold library, we further used a design of experiments approach to probe the parameter space and perform detailed analysis on chemical composition-scaffold properties as well as scaffold properties-cell behavior correlations. Our findings showed that hydrogel properties such as modulus, swelling, pore size, and permeability, strongly depended on total polymer concentration and not on the GelMA fraction. GelMA significantly influenced cell spreading, while addition of any amount of PEGDA delayed cell spreading significantly. We suggest that such analysis will broaden the utility of the GelMA/PEGDA hydrogels, presenting a versatile platform for mechanosensing research in 3D environments.

Finding synchronization state of higher-order motif networks by dynamic learning

Synchronization of action potentials is one of the important phenomena for neural networks to achieve biological functions. How to find the optimal parameter space of neural networks in a synchronized state is of great significance for studying their synchronization. This study explores the parameter space of triplet motif-based higher-order networks in a synchronized state through the dynamic learning of synchronization (DLS) technique, which dynamically modulates the connection weights between motifs to alter their firing patterns. Our study delves into regular, Erdós-Rënyi random graphs, small-world, and scale-free networks, emphasizing the high-order motif interactions that characterize these networks. Our key findings indicate that the DLS technique successfully promotes synchronization within high-order motif networks with various connection patterns, although the degree of synchronization in networks where motifs are interconnected by chemical synapses is slightly weaker than those interconnected by electrical synapses. Additionally, we demonstrate the pattern of weight changes during the regulation of network firing states by DLS, finding that the evolution of weight distributions correlates with the network's topological structure. This work might provide new insights into complex network synchronization and lays the foundation for further exploration of using DLS technology to synchronize higher-order networks through external factors. Published by the American Physical Society 2024

Chromomagnetic dipole moments of light quarks in the Bestest Little Higgs Model

Abstract In this paper, we research into the anomalous Chromomagnetic Dipole Moment (CMDM), denoted as $\hat{\mu}_{q}^{BLHM}$, of the light quarks $q=(u, c, d, s, b)$ within the framework of the Bestest Little Higgs Model (BLHM) as an extension of the Standard Model (SM). Our investigation encompasses novel interactions among the light quarks, the heavy quark $B$, and the heavy bosons $(W^{\prime\pm}, H^{\pm}, \phi^{\pm}, \eta^{\pm})$, incorporating the extended Cabibbo-Kobayashi-Maskawa (CKM) matrix characteristic of the BLHM. We thoroughly explore the permissible parameter space, yielding a spectrum of CMDM values ranging from $10^{-10}$ to $10^{-3}$.Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. Article funded by SCOAP3 and published under licence by Chinese Physical Society and the Institute of High Energy Physics of the Chinese Academy of Science and the Institute of Modern Physics of the Chinese Academy of Sciences and IOP Publishing Ltd.

A simple model for multiple equilibria in ice-covered oceans

Abstract The existence of multiple equilibria (ice-covered and ice-free states) is explored using a set of coupled, nondimensional equations that describe the heat and salt balances in basins, such as the Arctic Ocean, that are subject to atmospheric forcing and two distinct water mass sources. Six nondimensional numbers describe the influences of: atmospheric cooling; evaporation minus precipitation; solar radiation; atmospheric temperature, diapycnal mixing, and the temperature contrast between the two water masses. It is shown that multiple equilibria resulting from the dependence of albedo on ice cover exists over a wide range of parameter space, especially so in the weak mixing limit. Multiple equilibria can also occur if diapycnal mixing increases to O(10−4 m2 s−1) or larger under ice-free conditions due to enhanced upward mixing of warm, salty water from below. Sensitivities to various forcing parameter are discussed.

Constructing model-agnostic likelihoods, a method for the reinterpretation of particle physics results

Experimental High Energy Physics has entered an era of precision measurements. However, measurements of many of the accessible processes assume that the final states’ underlying kinematic distribution is the same as the Standard Model prediction. This assumption introduces an implicit model-dependency into the measurement, rendering the reinterpretation of the experimental analysis complicated without reanalysing the underlying data. We present a novel reweighting method in order to perform reinterpretation of particle physics measurements. It makes use of reweighting the Standard Model templates according to kinematic signal distributions of alternative theoretical models, prior to performing the statistical analysis. The generality of this method allows us to perform statistical inference in the space of theoretical parameters, assuming different kinematic distributions, according to a beyond Standard Model prediction. We implement our method as an extension to the pyhf software and interface it with the EOS software, which allows us to perform flavor physics phenomenology studies. Furthermore, we argue that, beyond the pyhf or HistFactory likelihood specification, only minimal information is necessary to make a likelihood model-agnostic and hence easily reinterpretable. We showcase that publishing such likelihoods is crucial for a full exploitation of experimental results.