Search NASA⌕ Search

SEARCH · Search NASA

Results for “trigger algorithms”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 343 records · Page 19

Intelligent experiments through real-time AI: Fast Data Processing and Autonomous Detector Control for sPHENIX and future EIC detectors

This R&D project, initiated by the DOE Nuclear Physics AI-Machine Learning initiative in 2022, leverages AI to address data processing challenges in high-energy nuclear experiments (RHIC, LHC, and future EIC). Our focus is on developing a demonstrator for real-time processing of high-rate data streams from sPHENIX experiment tracking detectors. The limitations of a 15 kHz maximum trigger rate imposed by the calorimeters can be negated by intelligent use of streaming technology in the tracking system. The approach efficiently identifies low momentum rare heavy flavor events in high-rate p+p collisions (3MHz), using Graph Neural Network (GNN) and High Level Synthesis for Machine Learning (hls4ml). Success at sPHENIX promises immediate benefits, minimizing resources and accelerating the heavy-flavor measurements. The approach is transferable to other fields. For the EIC, we develop a DIS-electron tagger using Artificial Intelligence - Machine Learning (AI-ML) algorithms for real-time identification, showcasing the transformative potential of AI and FPGA technologies in high-energy nuclear and particle experiments real-time data processing pipelines.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Efficient Solution Methods for Large-scale Optimization Problems Constrained by Time-dependent Partial Differential Equations (Final Report)

The objectives of this project are to devise and analyze efficient solvers for several classes of PDE- constrained optimization problems. The project was initially focused on, but not restricted to time- dependent PDE-constraints such as parabolic equations, time-dependent fluid flows (Navier-Stokes equations), hyperbolic equations (Burgers equation, shallow water equations). Other constraints considered are PDEs with stochastic coefficients. Additional elements of interest are the handling of problems with control and/or state inequality constraints or non-differentiable regularization terms, and boundary controls. A direction of work not initially listed among the goals, which was triggered by conversations with DOE scientists from the Sandia National Labs, is to develop efficient solvers for optimization- based domain-decomposition PDE-solvers. Also targeted in the project are a number of technical issues necessary for facilitating the integration of the methods resulted from this project with existing high-performance packages developed at DOE national labs; this integration is expected to further enable the scientific community to use the novel methods. An example of such a task is the need to develop algebraic multigrid (AMG) versions of the algorithms. A list of accomplishments in several areas is detailed as well as a list of publications.

97 MATHEMATICS AND COMPUTING↗

Topological grain boundary segregation transitions

Engineering the structure of grain boundaries (GBs) by solute segregation is a promising strategy to tailor the properties of polycrystalline materials. Solute segregation triggering phase transitions at GBs has been suggested theoretically to offer different pathways to design interfaces, but an understanding of their intrinsic atomistic nature is missing. Here, we combined atomic resolution electron microscopy and atomistic simulations to discover that iron segregation to GBs in titanium stabilizes icosahedral units (“cages”) that form robust building blocks of distinct GB phases. Owing to their five-fold symmetry, the iron cages cluster and assemble into hierarchical GB phases characterized by a different number and arrangement of the constituent icosahedral units. Our advanced GB structure prediction algorithms and atomistic simulations validate the stability of these observed phases and the high excess of iron at the GB that is accommodated by the phase transitions.

36 MATERIALS SCIENCE↗

Parallelized Kalman-Filter-Based Reconstruction of Particle Tracks on Many-Core Processors and GPUs

For over a decade now, physical and energy constraints have limited clock speed improvements in commodity microprocessors. Instead, chipmakers have been pushed into producing lower-power, multi-core processors such as Graphical Processing Units (GPU), ARM CPUs, and Intel MICs. Broad-based efforts from manufacturers and developers have been devoted to making these processors user-friendly enough to perform general computations. However, extracting performance from a larger number of cores, as well as specialized vector or SIMD units, requires special care in algorithm design and code optimization. One of the most computationally challenging problems in high-energy particle experiments is finding and fitting the charged-particle tracks during event reconstruction. This is expected to become by far the dominant problem at the High-Luminosity Large Hadron Collider (HL-LHC), for example. Today the most common track finding methods are those based on the Kalman filter. Experience with Kalman techniques on real tracking detector systems has shown that they are robust and provide high physics performance. This is why they are currently in use at the LHC, both in the trigger and offine. Previously we reported on the significant parallel speedups that resulted from our investigations to adapt Kalman filters to track fitting and track building on Intel Xeon and Xeon Phi. Here, we discuss our progresses toward the understanding of these processors and the new developments to port the Kalman filter to NVIDIA GPUs.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Trigger-DAQ and Slow Controls Systems in the Mu2e Experiment

The muon campus program at Fermilab includes the Mu2e experiment that will search for a charged-lepton flavor violating processes where a negative muon converts into an electron in the field of an aluminum nucleus, improving by four orders of magnitude the search sensitivity reached so far. Mu2e's Trigger and Data Acquisition System (TDAQ) uses otsdaq as its solution. Developed at Fermilab, otsdaq uses the artdaq DAQ framework and art analysis framework, under the-hood, for event transfer, filtering, and processing. otsdaq is an online DAQ software suite with a focus on flexibility and scalability, while providing a multi-user, web-based, interface accessible through the Chrome or Firefox web browser. The detector Read Out Controller (ROC), from the tracker and calorimeter, stream out zero-suppressed data continuously to the Data Transfer Controller (DTC). Data is then read over the PCIe bus to a software filter algorithm that selects events which are finally combined with the data flux that comes froma Cosmic Ray Ve to System (CRV). A Detector Control System (DCS) for monitoring, controlling, alarming, and archiving has been developed using the Experimental Physics and Industrial Control System (EPICS) Open Source Platform. The DCS System has also been itegrated into otsdaq. The installation of the TDAQ and the DCS systems in the Mu2e building is planned for 2021-2022, and a prototype has been built at Fermilab's Feynman Computing Center. We report here on the developments and achievements of the integration of Mu2e's DCS system into the online otsdaq software.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Status and perspectives of the ICARUS experiment at the Fermilab Short Baseline Neutrino program

In this study, the ICARUS collaboration has employed the 760 t T600 detector in a successful three-year physics run at the underground LNGS laboratory, performing a sensitive search for LSND-like anomalous ν e appearance in the CERN Neutrino to Gran Sasso beam, which contributed to the constraints on the allowed neutrino oscillation parameters to a narrow region around Δm 2 ~ 1 eV 2 . After a significant overhaul at CERN, the T600 detector has been installed at Fermilab. In 2020 the cryogenic commissioning began with detector cool down, liquid Argon filling and recirculation. ICARUS then started its operation collecting the first neutrino events from the Booster Neutrino Beam (BNB) and the Neutrinos at the Main Injector (NuMI) beam off-axis, which were used to test the ICARUS event selection, reconstruction and analysis algorithms. ICARUS successfully completed its commissioning phase in June 2022, moving then to data taking for neutrino oscillation physics, aiming at first to either confirm or refute the claim by Neutrino-4 short-baseline reactor. ICARUS will also jointly search for evidence of sterile neutrinos together with the Short-Baseline Near Detector, within the Fermilab Short-Baseline Neutrino program experiment, and will perform measurements of neutrino cross sections with both beams and several Beyond Standard Model searches with the NuMI beam. In this paper, the main technical achievements of the ICARUS detector subsystems (Time Projection Chambers, Light Detection System, Cosmic Ray Tagger, Trigger and Data Acquisition) obtained with both BNB and NuMI neutrino beams during the commissioning phase, will be presented in terms of the overall detector performance and capability to select and reconstruct neutrino events.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Description and Performance of Track and Primary-Vertex Reconstruction with the CMS Tracker

A description is provided of the software algorithms developed for the CMS tracker both for reconstructing charged-particle trajectories in proton-proton interactions and for using the resulting tracks to estimate the positions of the LHC luminous region and individual primary-interaction vertices. Despite the very hostile environment at the LHC, the performance obtained with these algorithms is found to be excellent. For tbar t events under typical 2011 pileup conditions, the average track-reconstruction efficiency for promptly-produced charged particles with transverse momenta of p(T) > 0.9GeV is 94% for pseudorapidities of |η| < 0.9 and 85% for 0.9 < |η| < 2.5. The inefficiency is caused mainly by hadrons that undergo nuclear interactions in the tracker material. For isolated muons, the corresponding efficiencies are essentially 100%. For isolated muons of p(T) = 100GeV emitted at |η| < 1.4, the resolutions are approximately 2.8% in p(T), and respectively, 10μm and 30μm in the transverse and longitudinal impact parameters. The position resolution achieved for reconstructed primary vertices that correspond to interesting pp collisions is 10–12μm in each of the three spatial dimensions. The tracking and vertexing software is fast and flexible, and easily adaptable to other functions, such as fast tracking for the trigger, or dedicated tracking for electrons that takes into account bremsstrahlung.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

An optimization-based approach to tailor the mechanical response of soft metamaterials undergoing rate-dependent instabilities

An optimization-based design framework is proposed to tune the response of soft metamaterials involving both geometric instabilities and nonlinear viscoelastic material behavior. Designing the response of soft metamaterials to harness instabilities and undergo large, tailored configuration changes will enable advancements in soft robotics, shock and vibration mitigation, and flexible electronics. In line with the metamaterial concept, the response of these materials is governed to a large extent by the geometric and topological makeup of their small-scale features. However, the link between structure and response is less intuitive for soft metamaterials due to their reliance upon highly nonlinear responses triggered by geometric instabilities. This is further complicated by the effects of viscoelastic relaxation, which recent studies have shown to alter the emergence of instabilities in non-intuitive ways. Here, these effects are accounted for in our framework to achieve various design objectives, including tailored force–displacement response and maximized energy absorption from both geometric and material effects. To fully automate this process, it is essential to have a completely robust equation solver for forward problems involving instabilities and viscoelastic relaxation. We achieve this by casting the search for stable mechanical equilibrium — i.e. the forward problem — as a minimization problem and utilize a trust region algorithm to robustly handle instabilities and follow energetically-favorable equilibrium paths through critical points.

97 MATHEMATICS AND COMPUTING↗

Ultrafast time-resolved x-ray absorption spectroscopy of ionized urea and its dimer through ab initio nonadiabatic dynamics

Investigating the early dynamics of chemical systems following ionization is essential for our understanding of radiation damage. However, experimental as well as theoretical investigations are very challenging due to the complex nature of these processes. Time-resolved x-ray absorption spectroscopy on a femtosecond timescale, in combination with appropriate simulations, is able to provide crucial insights into the ultrafast processes that occur upon ionization due to its element-specific probing nature. In this theoretical study, we investigate the ultrafast dynamics of valence-ionized states of urea and its dimer employing Tully's fewest switches surface hopping approach using Koopmans' theorem to describe the ionized system. We demonstrate that following valence ionization through a pump pulse, the time-resolved x-ray absorption spectra at the carbon, nitrogen, and oxygen K-edges reveal rich insights into the dynamics. Excited states of the ionized system give rise to time-delayed blueshifts in the x-ray absorption spectra as a result of electronic relaxation dynamics through nonadiabatic transitions. Moreover, our statistical analysis reveals specific structural dynamics in the molecule that induce time-dependent changes in the spectra. For the urea monomer, we elucidate the possibility to trace effects of specific molecular vibrations in the time-resolved x-ray absorption spectra. For the urea dimer, where ionization triggers a proton transfer reaction, we show how the x-ray absorption spectra can reveal specific details on the progress of proton transfer.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Multifamily Envelope Leakage Model

“The cost for blower testing is high, because it is labor intensive, and it may disrupt occupants in multiple units. This high cost and disruption deter program participants, and dissuade them from pursuing energy improvements that would trigger air leakage testing, such as improvements to the building envelope.” This statement found in a 2012 report by Heschong Mahone Group for several California interests emphasizes the importance of reducing the cost and complexity of blower testing in multifamily buildings. Energy efficiency opportunities are being bypassed. The cost of single blower testing is on the order of $300. The cost for guarded blower door testing—the more appropriate test for assessing energy savings opportunities—could easily be six times that, and that’s only if you have the equipment and simultaneous access to multiple apartments. Thus, the proper test is simply not performed. This research seeks to provide an algorithm for predicting the guarded blower door test result based upon a single, total blower door test.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Flame acceleration and transition to detonation in a pre-/main-chamber combustion system

Numerical simulations are performed to study the mechanism of deflagration to detonation transition (DDT) in a pre-/main-chamber combustion system with a stoichiometric ethylene–oxygen mixture. A Godunov algorithm, fifth-order in space, and third-order in time, is used to solve the fully compressible Navier–Stokes equations on a dynamically adapting mesh. A single-step, calibrated chemical diffusive model described by Arrhenius kinetics is used for energy release and conservation between the fuel and the product. The two-dimensional simulation shows that a laminar flame grows in the pre-chamber and then develops into a jet flame as it passes through the orifice. A strong shock forms immediately ahead of the flame, reflecting off the walls and interacting with the flame front. The shock–flame interactions are crucial for the development of flame instabilities, which trigger the subsequent flame development. The DDT arises due to a shock-focusing mechanism, where multiple shocks collide at the flame front. A chemical explosive mode analysis (CEMA) criterion is developed to study the DDT ignition mode. Preliminary one-dimensional computations for a laminar propagating flame, a fast flame deflagration, and a Chapman–Jouguet detonation are conducted to demonstrate the validity of CEMA on the chemical-diffusive model, as well as to determine the proper conditioning value for CEMA diagnostic. The two-dimensional analysis with CEMA indicates that the DDT initiated by the shock-focusing mechanism can form a strong thermal expansion region at the flame front that features large positive eigenvalues for the chemical explosive mode and dominance of the local autoignition mode. Thus, the CEMA criterion proposed in this study provides a robust diagnostic for identifying autoignition-supported DDT, of which the emergence of excessive local autoignition mode is found to be a precursor. The effect of grid size, initial temperature, and orifice size are then evaluated, and results show that although the close-chamber DDT is highly stochastic, the detonation initiation mechanism remains robust.

42 ENGINEERING↗

The Event Detection System in the NEXT-White Detector

This article describes the event detection system of the NEXT-White detector, a 5 kg high pressure xenon TPC with electroluminescent amplification, located in the Laboratorio Subterráneo de Canfranc (LSC), Spain. The detector is based on a plane of photomultipliers (PMTs) for energy measurements and a silicon photomultiplier (SiPM) tracking plane for offline topological event filtering. The event detection system, based on the SRS-ATCA data acquisition system developed in the framework of the CERN RD51 collaboration, has been designed to detect multiple events based on online PMT signal energy measurements and a coincidence-detection algorithm. Implemented on FPGA, the system has been successfully running and evolving during NEXT-White operation. The event detection system brings some relevant and new functionalities in the field. A distributed double event processor has been implemented to detect simultaneously two different types of events thus allowing simultaneous calibration and physics runs. This special feature provides constant monitoring of the detector conditions, being especially relevant to the lifetime and geometrical map computations which are needed to correct high-energy physics events. Other features, like primary scintillation event rejection, or a double buffer associated with the type of event being searched, help reduce the unnecessary data throughput thus minimizing dead time and improving trigger efficiency.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Using Cosmic Ray Muons to Assess Geological Characteristics in the Subsurface

Cosmic rays are energetic nuclei and elementary particles that originate from stars and intergalactic events. The interaction of these particles with the upper atmosphere produces a wide range of secondary particles that reach the surface of the earth, of which muons are the most prominent. With enough energy, muons can travel up to a few kilometers beneath the surface of the earth before being stopped completely. The terrestrial muon flux profile and associated zenith angle can be utilized to determine geological characteristics of a location (e.g., rock overburden and density) without having to use conventional methods such as boreholes. This work uses a low-power plastic scintillator-based muon detection system as a prototype for this non-destructive geological assay methodology. Four custom designed 102 cm x 51 cm x 5 cm plastic scintillation panels are used to realize two orthogonal detection planes. Optical photons from each scintillation panel are read using OnSemi J-Series 4x4 silicon photomultiplier (SiPM) arrays in conjunction with preamplifiers. Simultaneous triggers between detectors from two planes indicate a coincidence event which is recorded using the QuarkNet data acquisition system (DAQ) from Fermi National Accelerator Laboratory. A custom detector holder was designed to securely mount the detection system and rotate the panels along the zenith to collect data at variable angles. In order to quantify the systematic uncertainties associated with the detector, such as energy depositions and angular resolution of the detector design, a Monte Carlo (MC) simulation using Geant4 is being developed. Cosmic ray flux prediction will be included in the project by adding the CORSIKA MC code to the simulation toolchain. Simulated and experimental data will drive the development and validation of a reconstruction algorithm that, upon completion, is expected to predict average overburden and rock density. Extended detector exposure to muons can be used as a means to understand changes in the surrounding environment like rock porosity. On the experimental front, muons will initially be measured at the surface, establishing the baseline flux. This is followed by recording the muon flux at variable depths and zenith angles, where the data will be used by the reconstruction algorithm to predict the overburden. The result will be benchmarked against geological surveys. The measured flux data will also be used to benchmark independent and established models. Successful proof-of-concept demonstration of this technology can open doors for long term non-invasive geological monitoring. The detector design, experimental methodology, and the benchmarking efforts are detailed in this work.

Gadey, Harish Reddy↗

Simulation-driven optimization of high-order meshes in ALE hydrodynamics

Here we propose tools for high-order mesh optimization and demonstrate their benefits in the context of multi-material Arbitrary Lagrangian-Eulerian (ALE) compressible shock hydrodynamic applications. The mesh optimization process is driven by information provided by the simulation which uses the optimized mesh, such as shock positions, material regions, known error estimates, etc. These simulation features are usually represented discretely, for instance, as finite element functions on the Lagrangian mesh. The discrete nature of the input is critical for the practical applicability of the algorithms we propose and distinguishes this work from approaches that strictly require analytical information. Our methods are based on node movement through a high-order extension of the Target-Matrix Optimization Paradigm (TMOP). The proposed formulation is fully algebraic and relies only on local Jacobian matrices, so it is applicable to all types of mesh elements, in 2D and 3D, and any order of the mesh. We discuss the notions of constructing adaptive target matrices and obtaining their derivatives, reconstructing discrete data in intermediate meshes, node limiting that enables improvement of global mesh quality while preserving space-dependent local mesh features, and appropriate normalization of the objective function. The adaptivity methods are combined with automatic ALE triggers that can provide robustness of the mesh evolution and avoid excessive remap procedures. The benefits of the new high-order TMOP technology are illustrated on several simulations performed in the high-order ALE application BLAST.

97 MATHEMATICS AND COMPUTING↗

Distributed Inference with Sparse and Quantized Communication

Here, we consider the problem of distributed inference where agents in a network observe a stream of private signals generated by an unknown state, and aim to uniquely identify this state from a finite set of hypotheses. We focus on scenarios where communication between agents is costly, and takes place over channels with finite bandwidth. To reduce the frequency of communication, we develop a novel event-triggered distributed learning rule that is based on the principle of diffusing low beliefs on each false hypothesis. Building on this principle, we design a trigger condition under which an agent broadcasts only those components of its belief vector that have adequate innovation, to only those neighbors that require such information. We prove that our rule guarantees convergence to the true state exponentially fast almost surely despite sparse communication, and that it has the potential to significantly reduce information flow from uninformative agents to informative agents. Next, to deal with finite-precision communication channels, we propose a distributed learning rule that leverages the idea of adaptive quantization. We show that by sequentially refining the range of the quantizers, every agent can learn the truth exponentially fast almost surely, while using just 1 bit to encode its belief on each hypothesis. For both our proposed algorithms, we rigorously characterize the trade-offs between communication-efficiency and the learning rate.

42 ENGINEERING↗

EQUILIBRIUM DISKS, MAGNETOROTATIONAL INSTABILITY MODE EXCITATION, AND STEADY-STATE TURBULENCE IN GLOBAL ACCRETION DISK SIMULATIONS

Global three-dimensional magnetohydrodynamic (MHD) simulations of turbulent accretion disks are presented which start from fully equilibrium initial conditions in which the magnetic forces are accounted for and the induction equation is satisfied. The local linear theory of the magnetorotational instability (MRI) is used as a predictor of the growth of magnetic field perturbations in the global simulations. The linear growth estimates and global simulations diverge when nonlinear motions-perhaps triggered by the onset of turbulence-upset the velocity perturbations used to excite the MRI. The saturated state is found to be independent of the initially excited MRI mode, showing that once the disk has expelled the initially net flux field and settled into quasi-periodic oscillations in the toroidal magnetic flux, the dynamo cycle regulates the global saturation stress level. Furthermore, time-averaged measures of converged turbulence, such as the ratio of magnetic energies, are found to be in agreement with previous works. In particular, the globally averaged stress normalized to the gas pressure <{alpha}{sub P}>bar = 0.034, with notably higher values achieved for simulations with higher azimuthal resolution. Supplementary tests are performed using different numerical algorithms and resolutions. Convergence with resolution during the initial linear MRI growth phase is found for 23-35 cells per scale height (in the vertical direction).

79 ASTRONOMY AND ASTROPHYSICS↗

Position-probability-sampled Monte Carlo calculation of VMAT, 3DCRT, step-shoot IMRT, and helical tomotherapy dose distributions using BEAMnrc/DOSXYZnrc

Purpose: The commercial release of volumetric modulated arc therapy techniques using a conventional linear accelerator and the growing number of helical tomotherapy users have triggered renewed interest in dose verification methods, and also in tools for exploring the impact of machine tolerance and patient motion on dose distributions without the need to approximate time-varying parameters such as gantry position, MLC leaf motion, or patient motion. To this end we have developed a Monte Carlo-based calculation method capable of simulating a wide variety of treatment techniques without the need to resort to discretization approximations. Methods: The ability to perform complete position-probability-sampled Monte Carlo dose calculations was implemented in the BEAMnrc/DOSXZYnrc user codes of EGSnrc. The method includes full accelerator head simulations of our tomotherapy and Elekta linacs, and a realistic representation of continous motion via the sampling of a time variable. The functionality of this algorithm was tested via comparisons with both measurements and treatment planning dose distributions for four types of treatment techniques: 3D conformal, step-shoot intensity modulated radiation therapy, helical tomotherapy, and volumetric modulated arc therapy. Results: For static fields, the absolute dose agreement between the EGSnrc Monte Carlo calculations and measurements is within 2%/1 mm. Absolute dose agreement between Monte Carlo calculations and treatment planning system for the four different treatment techniques is within 3%/3 mm. Discrepancies with the tomotherapy TPS on the order of 10%/5 mm were observed for the extreme example of a small target located 15 cm off-axis and planned with a low modulation factor. The increase in simulation time associated with using position-probability sampling, as opposed to the discretization approach, was less than 2% in most cases. Conclusions: A single Monte Carlo simulation method can be used to calculate patient dose distribution for various types of treatment techniques delivered with either tomotherapy or a conventional linac. The method simplifies the simulation process, improves dose calculation accuracy, and involves an acceptably small change in computation time.

60 APPLIED LIFE SCIENCES↗

Bias-Variance Trade-Off in Physics-Informed Neural Networks with Randomized Smoothing for High-Dimensional PDEs

Physics-Informed Neural Networks (PINNs) have triggered a paradigm shift in scientific computing, leveraging mesh-free properties and robust approximation capabilities. While proving effective for low-dimensional partial differential equations (PDEs), the computational cost of PINNs remains a hurdle in high-dimensional scenarios. This is particularly pronounced when computing high-order and high-dimensional derivatives in the physics-informed loss. Randomized Smoothing PINN (RS-PINN) introduces Gaussian noise for stochastic smoothing of the original neural net model, enabling the use of Monte Carlo methods for derivative approximation, which eliminates the need for costly automatic differentiation. Despite its computational efficiency, especially in the approximation of high-dimensional derivatives, RS-PINN introduces biases in both loss and gradients, negatively impacting convergence, especially when coupled with stochastic gradient descent (SGD) algorithms. We present a comprehensive analysis of biases in RS-PINN, attributing them to the nonlinearity of the Mean Squared Error (MSE) loss as well as the intrinsic nonlinearity of the PDE itself. We propose tailored bias correction techniques, delineating their application based on the order of PDE nonlinearity. The derivation of an unbiased RS-PINN allows for a detailed examination of its advantages and disadvantages compared to the biased version. Specifically, the biased version has a lower variance and runs faster than the unbiased version, but it is less accurate due to the bias. To optimize the bias-variance trade-off, we combine the two approaches in a hybrid method that balances the rapid convergence of the biased version with the high accuracy of the unbiased version. In addition to methodological contributions, we present an enhanced implementation of RS-PINN. Extensive experiments on diverse high-dimensional PDEs, including Fokker-Planck, Hamilton-Jacobi-Bellman (HJB), viscous Burgers’, Allen-Cahn, and Sine-Gordon equations, illustrate the bias-variance trade-off and highlight the effectiveness of the hybrid RS-PINN. Empirical guidelines are provided for selecting biased, unbiased, or hybrid versions, depending on the dimensionality and nonlinearity of the specific PDE problem.

97 MATHEMATICS AND COMPUTING↗