The data-driven machine learning paradigm typically requires high-quality,large-scale datasets for training neural networks,which are often unavailable in many scientific and engineering applications.Integrating physi...The data-driven machine learning paradigm typically requires high-quality,large-scale datasets for training neural networks,which are often unavailable in many scientific and engineering applications.Integrating physics equations into machine learning models,either fully or partially,can mitigate these data requirements and improve generalizability;however,such approaches frequently rely on differentiable programming frameworks.This ability poses significant challenges when legacy or commercial numerical solvers,which are often nondifferentiable and difficult to modify without introducing code changes,are integrated.This work addresses these challenges by leveraging the mini-batching iterative ensemble Kalman inversion(EKI)algorithm as a gradientfree training framework for hybrid neural models.The use of stochastic mini-batching significantly enhances the computational efficiency and convergence of EKI,making it well-suited for high-dimensional learning problems.The proposed method is demonstrated for modeling a fiber-reinforced composite plate,where heterogeneous local constitutive laws are parameterized by a trainable neural network embedded within the FEniCS finite element solver.Using the displacement field as indirect data,the hybrid neural FEM solver successfully predicts deformations by learning the local constitutive laws,even for unseen fiber volume fraction distributions and varying test loading conditions.These results demonstrate the effectiveness of iterative EKI in training hybrid neural models with non-differentiable components,paving the way for broader adoption of hybrid neural models in scientific and engineering applications.展开更多
Non-coding variants in the human genome significantly influence human traits and complex diseases via their regulation and modification effects.Hence,an increasing number of computational methods are developed to pred...Non-coding variants in the human genome significantly influence human traits and complex diseases via their regulation and modification effects.Hence,an increasing number of computational methods are developed to predict the effects of variants in human non-coding sequences.However,it is difficult for inexperienced users to select appropriate computational methods from dozens of available methods.To solve this issue,we assessed 12 performance metrics of 24 methods on four independent non-coding variant benchmark datasets:(1)rare germline variants from clinical relevant sequence variants(ClinVar),(2)rare somatic variants from Catalogue Of Somatic Mutations In Cancer(COSMIC),(3)common regulatory variants from curated expression quantitative trait locus(eQTL)data,and(4)disease-associated common variants from curated genomewide association studies(GWAS).All 24 tested methods performed differently under various conditions,indicating varying strengths and weaknesses under different scenarios.Importantly,the performance of existing methods was acceptable for rare germline variants from ClinVar with the area under the receiver operating characteristic curve(AUROC)of 0.4481–0.8033 and poor for rare somatic variants from COSMIC(AUROC=0.4984–0.7131),common regulatory variants from curated eQTL data(AUROC=0.4837–0.6472),and disease-associated common variants from curated GWAS(AUROC=0.4766–0.5188).We also compared the prediction performance of 24 methods for non-coding de novo mutations in autism spectrum disorder,and found that the combined annotation-dependent depletion(CADD)and context-dependent tolerance score(CDTS)methods showed better performance.Summarily,we assessed the performance of 24 computational methods under diverse scenarios,providing preliminary advice for proper tool selection and guiding the development of new techniques in interpreting non-coding variants.展开更多
Phonon Boltzmann transport equation(BTE)is a key tool for modeling multiscale phonon transport,which is critical to the thermal management of miniaturized integrated circuits,but assumptions about the system temperatu...Phonon Boltzmann transport equation(BTE)is a key tool for modeling multiscale phonon transport,which is critical to the thermal management of miniaturized integrated circuits,but assumptions about the system temperatures(i.e.,small temperature gradients)are usually made to ensure that it is computationally tractable.To include the effects of large temperature non-equilibrium,we demonstrate a data-free deep learning scheme,physics-informed neural network(PINN),for solving stationary,mode-resolved phonon BTE with arbitrary temperature gradients.This scheme uses the temperature-dependent phonon relaxation times and learns the solutions in parameterized spaces with both length scale and temperature gradient treated as input variables.Numerical experiments suggest that the proposed PINN can accurately predict phonon transport(from 1D to 3D)under arbitrary temperature gradients.Moreover,the proposed scheme shows great promise in simulating device-level phonon heat conduction efficiently and can be potentially used for thermal design.展开更多
Chemical vapor infiltration(CVI)is a widely adopted manufacturing technique used in producing carbon-carbon and carbon-silicon carbide composites.These materials are especially valued in the aerospace and automotive i...Chemical vapor infiltration(CVI)is a widely adopted manufacturing technique used in producing carbon-carbon and carbon-silicon carbide composites.These materials are especially valued in the aerospace and automotive industries for their robust strength and lightweight characteristics.The densification process during CVI critically influences the final performance,quality,and consistency of these composite materials.Experimentally optimizing the CVI processes is challenging due to the long experimental time and large optimization space.To address these challenges,this work takes a modeling-centric approach.Due to the complexities and limited experimental data of the isothermal CVI densification process,we have developed a data-driven predictive model using the physicsintegrated neural differentiable(PiNDiff)modeling framework.An uncertainty quantification feature has been embedded within the PiNDiff method,bolstering the model’s reliability and robustness.Through comprehensive numerical experiments involving both synthetic and real-world manufacturing data,the proposed method showcases its capability in modeling densification during the CVI process.This research highlights the potential of the PiNDiff framework as an instrumental tool for advancing our understanding,simulation,and optimization of the CVI manufacturing process,particularly when faced with sparse data and an incomplete description of the underlying physics.展开更多
The phonon Boltzmann transport equation(BTE)is a powerful tool for modeling and understanding micro-/nanoscale thermal transport in solids,where Fourier’s law can fail due to non-diffusive effect when the characteris...The phonon Boltzmann transport equation(BTE)is a powerful tool for modeling and understanding micro-/nanoscale thermal transport in solids,where Fourier’s law can fail due to non-diffusive effect when the characteristic length/time is comparable to the phonon mean free path/relaxation time.However,numerically solving phonon BTE can be computationally costly due to its high dimensionality,especially when considering mode-resolved phonon properties and time dependency.In this work,we demonstrate the effectiveness of physics-informed neural networks(PINNs)in solving time-dependent mode-resolved phonon BTE.The PINNs are trained by minimizing the residual of the governing equations,and boundary/initial conditions to predict phonon energy distributions,without the need for any labeled training data.The results obtained using the PINN framework demonstrate excellent agreement with analytical and numerical solutions.Moreover,after offline training,the PINNs can be utilized for online evaluation of transient heat conduction,providing instantaneous results,such as temperature distribution.It is worth noting that the training can be carried out in a parametric setting,allowing the trained model to predict phonon transport in arbitrary values in the parameter space,such as the characteristic length.This efficient and accurate method makes it a promising tool for practical applications such as the thermal management design of microelectronics.展开更多
Quantum algorithms are emerging tools in the design of functional materials due to their powerful solution space search capability.How to balance the high price of quantum computing resources and the growing computing...Quantum algorithms are emerging tools in the design of functional materials due to their powerful solution space search capability.How to balance the high price of quantum computing resources and the growing computing needs has become an urgent problem to be solved.We propose a novel optimization strategy based on an active learning scheme that combines the Quantum-inspired Genetic Algorithm(QGA)with machine learning surrogate model regression.Using Random Forests as the surrogate model circumvents the time-consuming physical modeling or experiments,thereby improving the optimization efficiency.QGA,a genetic algorithm embedded with quantum mechanics,combines the advantages of quantum computing and genetic algorithms,enabling faster and more robust convergence to the optimum.Using the design of planar multilayer photonic structures for transparent radiative cooling as a testbed,we show superiority of our algorithm over the classical genetic algorithm(CGA).Additionally,we show the precision advantage of the Random Forest(RF)model as a flexible surrogate model,which relaxes the constraints on the type of surrogate model that can be used in other quantum computing optimization algorithms(e.g.,quantum annealing needs Ising model as a surrogate).展开更多
Polymers play an integral role in various applications,from everyday use to advanced technologies.In the era of machine learning(ML),polymer informatics has become a vital field for efficiently designing and developin...Polymers play an integral role in various applications,from everyday use to advanced technologies.In the era of machine learning(ML),polymer informatics has become a vital field for efficiently designing and developing polymeric materials.However,the focus of polymer informatics has predominantly centered on single-component polymers,leaving the vast chemical space of polymer blends relatively unexplored.This study employs a high-throughput molecular dynamics(MD)simulation combined with active learning(AL)to uncover polymer blends with enhanced thermal conductivity(TC)compared to the constituent single-component polymers.Initially,the TC of about 600 amorphous single-component polymers and 200 amorphous polymer blends with varying blending ratios are determined through MD simulations.The optimal representation method for polymer blends is identified,which involves a weighted sum approach that extends existing polymer representation from single-component polymers to polymer blends.An AL framework,combining MD simulation and ML,is employed to explore the TC of approximately 550,000 unlabeled polymer blends.The AL framework proves highly effective in accelerating the discovery of high-performance polymer blends for thermal transport.Additionally,we delve into the relationship between TC,radius of gyration(Rg),and hydrogen bonding,highlighting the roles of inter-and intra-chain interactions in thermal transport in amorphous polymer blends.A significant positive association between TC and Rg improvement and an indirect contribution from H-bond interaction to TC enhancement are revealed through a log-linear model and an odds ratio calculation,emphasizing the impact of increasingRg and H-bond interactions on enhancing polymer blend TC.展开更多
基金supported by the Air Force Office of Scientific Research(AFOSR),United States of America(Grant No.FA9550-22-1-0065).
文摘The data-driven machine learning paradigm typically requires high-quality,large-scale datasets for training neural networks,which are often unavailable in many scientific and engineering applications.Integrating physics equations into machine learning models,either fully or partially,can mitigate these data requirements and improve generalizability;however,such approaches frequently rely on differentiable programming frameworks.This ability poses significant challenges when legacy or commercial numerical solvers,which are often nondifferentiable and difficult to modify without introducing code changes,are integrated.This work addresses these challenges by leveraging the mini-batching iterative ensemble Kalman inversion(EKI)algorithm as a gradientfree training framework for hybrid neural models.The use of stochastic mini-batching significantly enhances the computational efficiency and convergence of EKI,making it well-suited for high-dimensional learning problems.The proposed method is demonstrated for modeling a fiber-reinforced composite plate,where heterogeneous local constitutive laws are parameterized by a trainable neural network embedded within the FEniCS finite element solver.Using the displacement field as indirect data,the hybrid neural FEM solver successfully predicts deformations by learning the local constitutive laws,even for unseen fiber volume fraction distributions and varying test loading conditions.These results demonstrate the effectiveness of iterative EKI in training hybrid neural models with non-differentiable components,paving the way for broader adoption of hybrid neural models in scientific and engineering applications.
基金supported by the National Natural Science Foundation of China(Grant No.81801133 to JL)the Young Elite Scientist Sponsorship Program by China Association for Science and Technology(Grant No.2018QNRC001 to JL)+2 种基金the Innovation-Driven Project of Central South University,China(Grant No.20180033040004 to JL)the Natural Science Foundation for Young Scientists of Hunan Province,China(Grant No.2019JJ50974 to GZ)the Natural Science Foundation of Hunan Province for outstanding Young Scholars,China(Grant No.2020JJ3059 to JL).
文摘Non-coding variants in the human genome significantly influence human traits and complex diseases via their regulation and modification effects.Hence,an increasing number of computational methods are developed to predict the effects of variants in human non-coding sequences.However,it is difficult for inexperienced users to select appropriate computational methods from dozens of available methods.To solve this issue,we assessed 12 performance metrics of 24 methods on four independent non-coding variant benchmark datasets:(1)rare germline variants from clinical relevant sequence variants(ClinVar),(2)rare somatic variants from Catalogue Of Somatic Mutations In Cancer(COSMIC),(3)common regulatory variants from curated expression quantitative trait locus(eQTL)data,and(4)disease-associated common variants from curated genomewide association studies(GWAS).All 24 tested methods performed differently under various conditions,indicating varying strengths and weaknesses under different scenarios.Importantly,the performance of existing methods was acceptable for rare germline variants from ClinVar with the area under the receiver operating characteristic curve(AUROC)of 0.4481–0.8033 and poor for rare somatic variants from COSMIC(AUROC=0.4984–0.7131),common regulatory variants from curated eQTL data(AUROC=0.4837–0.6472),and disease-associated common variants from curated GWAS(AUROC=0.4766–0.5188).We also compared the prediction performance of 24 methods for non-coding de novo mutations in autism spectrum disorder,and found that the combined annotation-dependent depletion(CADD)and context-dependent tolerance score(CDTS)methods showed better performance.Summarily,we assessed the performance of 24 computational methods under diverse scenarios,providing preliminary advice for proper tool selection and guiding the development of new techniques in interpreting non-coding variants.
基金The authors would like to thank ONR MURI(N00014-18-1-2429)for the financial support.The simulations are supported by the Notre Dame Center for Research ComputingNSF through the eXtreme Science and Engineering Discovery Environment(XSEDE)computing resources provided by Texas Advanced Computing Center(TACC)Stampede II under grant number TG-CTS100078This work is also supported by the National Research Foundation of Korea(NRF)grant funded by the Korea government(MSIT)(No.NRF-2021R1C1C1006251).
文摘Phonon Boltzmann transport equation(BTE)is a key tool for modeling multiscale phonon transport,which is critical to the thermal management of miniaturized integrated circuits,but assumptions about the system temperatures(i.e.,small temperature gradients)are usually made to ensure that it is computationally tractable.To include the effects of large temperature non-equilibrium,we demonstrate a data-free deep learning scheme,physics-informed neural network(PINN),for solving stationary,mode-resolved phonon BTE with arbitrary temperature gradients.This scheme uses the temperature-dependent phonon relaxation times and learns the solutions in parameterized spaces with both length scale and temperature gradient treated as input variables.Numerical experiments suggest that the proposed PINN can accurately predict phonon transport(from 1D to 3D)under arbitrary temperature gradients.Moreover,the proposed scheme shows great promise in simulating device-level phonon heat conduction efficiently and can be potentially used for thermal design.
基金The authors would like to acknowledge the funds from the Air Force Office of Scientific Research(AFOSR),United States of America,under award number FA9550-22-1-0065J.X.W.would also like to acknowledge the funding support from the Office of Naval Research under award number N00014-23-1-2071the National Science Foundation under award number OAC-2047127 in supporting this study.
文摘Chemical vapor infiltration(CVI)is a widely adopted manufacturing technique used in producing carbon-carbon and carbon-silicon carbide composites.These materials are especially valued in the aerospace and automotive industries for their robust strength and lightweight characteristics.The densification process during CVI critically influences the final performance,quality,and consistency of these composite materials.Experimentally optimizing the CVI processes is challenging due to the long experimental time and large optimization space.To address these challenges,this work takes a modeling-centric approach.Due to the complexities and limited experimental data of the isothermal CVI densification process,we have developed a data-driven predictive model using the physicsintegrated neural differentiable(PiNDiff)modeling framework.An uncertainty quantification feature has been embedded within the PiNDiff method,bolstering the model’s reliability and robustness.Through comprehensive numerical experiments involving both synthetic and real-world manufacturing data,the proposed method showcases its capability in modeling densification during the CVI process.This research highlights the potential of the PiNDiff framework as an instrumental tool for advancing our understanding,simulation,and optimization of the CVI manufacturing process,particularly when faced with sparse data and an incomplete description of the underlying physics.
基金The authors would like to thank ONR MURI(N00014-18-1-2429)and DARPA(HR00112390112)for the financial supportThe simulations are supported by the Notre Dame Center for Research Computing,and NSF through the eXtreme Science and Engineering Discovery Environment(XSEDE)computing resources provided by Texas Advanced Computing Center(TACC)Stampede II under grant number TG-CTS100078.
文摘The phonon Boltzmann transport equation(BTE)is a powerful tool for modeling and understanding micro-/nanoscale thermal transport in solids,where Fourier’s law can fail due to non-diffusive effect when the characteristic length/time is comparable to the phonon mean free path/relaxation time.However,numerically solving phonon BTE can be computationally costly due to its high dimensionality,especially when considering mode-resolved phonon properties and time dependency.In this work,we demonstrate the effectiveness of physics-informed neural networks(PINNs)in solving time-dependent mode-resolved phonon BTE.The PINNs are trained by minimizing the residual of the governing equations,and boundary/initial conditions to predict phonon energy distributions,without the need for any labeled training data.The results obtained using the PINN framework demonstrate excellent agreement with analytical and numerical solutions.Moreover,after offline training,the PINNs can be utilized for online evaluation of transient heat conduction,providing instantaneous results,such as temperature distribution.It is worth noting that the training can be carried out in a parametric setting,allowing the trained model to predict phonon transport in arbitrary values in the parameter space,such as the characteristic length.This efficient and accurate method makes it a promising tool for practical applications such as the thermal management design of microelectronics.
基金supported by the Quantum Computing Based on Quantum Advantage Challenge Research(RS-2023-00255442)through the National Research Foundation of Korea(NRF)funded by the Korean government(Ministry of Science and ICT(MSIT))This research also used resources of the Oak Ridge Leadership Computing Facility at the Oak Ridge National Laboratory,which is supported by the Office of Science of the U.S.Department of Energy under Contract No.DE-AC05-00OR22725.
文摘Quantum algorithms are emerging tools in the design of functional materials due to their powerful solution space search capability.How to balance the high price of quantum computing resources and the growing computing needs has become an urgent problem to be solved.We propose a novel optimization strategy based on an active learning scheme that combines the Quantum-inspired Genetic Algorithm(QGA)with machine learning surrogate model regression.Using Random Forests as the surrogate model circumvents the time-consuming physical modeling or experiments,thereby improving the optimization efficiency.QGA,a genetic algorithm embedded with quantum mechanics,combines the advantages of quantum computing and genetic algorithms,enabling faster and more robust convergence to the optimum.Using the design of planar multilayer photonic structures for transparent radiative cooling as a testbed,we show superiority of our algorithm over the classical genetic algorithm(CGA).Additionally,we show the precision advantage of the Random Forest(RF)model as a flexible surrogate model,which relaxes the constraints on the type of surrogate model that can be used in other quantum computing optimization algorithms(e.g.,quantum annealing needs Ising model as a surrogate).
基金funding support from the U.S.National Science Foundation(CDSE-2102592).
文摘Polymers play an integral role in various applications,from everyday use to advanced technologies.In the era of machine learning(ML),polymer informatics has become a vital field for efficiently designing and developing polymeric materials.However,the focus of polymer informatics has predominantly centered on single-component polymers,leaving the vast chemical space of polymer blends relatively unexplored.This study employs a high-throughput molecular dynamics(MD)simulation combined with active learning(AL)to uncover polymer blends with enhanced thermal conductivity(TC)compared to the constituent single-component polymers.Initially,the TC of about 600 amorphous single-component polymers and 200 amorphous polymer blends with varying blending ratios are determined through MD simulations.The optimal representation method for polymer blends is identified,which involves a weighted sum approach that extends existing polymer representation from single-component polymers to polymer blends.An AL framework,combining MD simulation and ML,is employed to explore the TC of approximately 550,000 unlabeled polymer blends.The AL framework proves highly effective in accelerating the discovery of high-performance polymer blends for thermal transport.Additionally,we delve into the relationship between TC,radius of gyration(Rg),and hydrogen bonding,highlighting the roles of inter-and intra-chain interactions in thermal transport in amorphous polymer blends.A significant positive association between TC and Rg improvement and an indirect contribution from H-bond interaction to TC enhancement are revealed through a log-linear model and an odds ratio calculation,emphasizing the impact of increasingRg and H-bond interactions on enhancing polymer blend TC.
