检索结果-内蒙古大学图书馆

An integrated framework for accelerating reactive flow simulation using GPU and machine learning models

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Mao, Runze Wang, Yingrui Zhang, Min Li, Han Xu, Jiayang Dong, Xinyu Zhang, Yan Chen, Zhi X. State Key Laboratory of Turbulence and Complex Systems Aeronautics and Astronautics College of Engineering Peking University Beijing100871 China Beijing100080 China Shanghai SenseTime Intelligent Technology Co. LTD Shanghai200233 China CAEP Software Center for High Performance Numerical Simulation Beijing100088 China Institute of Applied Physics and Computational Mathematics Beijing100088 China

Recent progress in artificial intelligence (AI) and high-performance computing (HPC) have brought potentially game-changing opportunities in accelerating reactive flow simulations. In this study, we introduce an open-source computational fluid dynamics (CFD) framework that integrates the strengths of machine learning (ML) and graphics processing unit (GPU) to demonstrate their combined capability. Within this framework, all computational operations are solely executed on GPU, including ML-accelerated chemistry integration, fully-implicit solving of PDEs, and computation of thermal and transport properties, thereby eliminating the CPU-GPU memory copy overhead. Optimisations both within the kernel functions and during the kernel launch process are conducted to enhance computational performance. Strategies such as static data reorganisation and dynamic data allocation are adopted to reduce the GPU memory footprint. The computational performance is evaluated in two turbulent flame benchmarks using quasi-DNS and LES modelling, respectively. Remarkably, while maintaining a similar level of accuracy to the conventional CPU/CVODE-based solver, the GPU/ML-accelerated approach shows an overall speedup of over two orders of magnitude for both cases. This result highlights that high-fidelity turbulent combustion simulation with finite-rate chemistry that requires normally hundreds of CPUs can now be performed on portable devices such as laptops with a medium-end GPU. © 2023, CC BY.

关键词： Graphics processing unit

Encoding classical information into quantum resources

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Korzekwa, Kamil Puchala, Zbigniew Tomamichel, Marco Zyczkowski, Karol International Centre for Theory of Quantum Technologies University of Gdansk Gdańsk80-308 Poland Faculty of Physics Astronomy and Applied Computer Science Jagiellonian University Kraków30-348 Poland Institute of Theoretical and Applied Informatics Polish Academy of Sciences Gliwice44-100 Poland Centre for Quantum Technologies National University of Singapore Singapore117543 Singapore Department of Electrical and Computer Engineering National University of Singapore Singapore117583 Singapore Centre for Quantum Software and Information School of Software University of Technology Sydney SydneyNSW2007 Australia Center for Theoretical Physics Polish Academy of Sciences Warszawa02-668 Poland

We introduce and analyse the problem of encoding classical information into different resources of a quantum state. More precisely, we consider a general class of communication scenarios characterised by encoding operations that commute with a unique resource destroying map and leave free states invariant. Our motivating example is given by encoding information into coherences of a quantum system with respect to a fixed basis (with unitaries diagonal in that basis as encodings and the decoherence channel as a resource destroying map), but the generality of the framework allows us to explore applications ranging from super-dense coding to thermodynamics. For any state, we find that the number of messages that can be encoded into it using such operations in a one-shot scenario is upper bounded in terms of the information spectrum relative entropy between the given state and its version with erased resources. Furthermore, if the resource destroying map is the twirling channel over some unitary group, we find matching one-shot lower bounds as well. In the asymptotic setting where we encode into many copies of the resource state, our bounds yield an operational interpretation of resource monotones such as the relative entropy of coherence and its corresponding relative entropy variance. Copyright © 2019, The Authors. All rights reserved.

关键词： Quantum optics

Concurrent learning scheme for crystal structure prediction

学校读者我要写书评

暂无评论

Physical Review B 2024年第9期109卷 094117-094117页

作者： Zhenyu Wang Xiaoyang Wang Xiaoshan Luo Pengyue Gao Ying Sun Jian Lv Han Wang Yanchao Wang Yanming Ma Key Laboratory of Material Simulation Methods and Software of Ministry of Education College of Physics Jilin University Changchun 130012 People's Republic of China International Center of Future Science Jilin University Changchun 130012 People's Republic of China Laboratory of Computational Physics Institute of Applied Physics and Computational Mathematics Fenghao East Road 2 Beijing 100094 People's Republic of China HEDPS CAPT College of Engineering Peking University Beijing 100871 People's Republic of China

Crystal structure prediction (CSP) and machine learning potential (MLP) are two fundamental methods for modern computational material discovery. While the former aims at efficient sampling of the potential energy surface (PES) for discovering new materials, the latter focuses on reproducing the PES to accelerate various atomic simulation tasks. In this work, we combine the two methods within a concurrent learning framework in an effort to generate efficient MLP models for accelerating CSP. The proposed scheme explores the PES through the swarm-intelligence calypso method, labels the most representative structures with quantum mechanical calculations, and learns the PES through a deep potential (DP) model. The process proceeds in an iterative, computationally efficient, and automated manner, leading to the collection of a most compact reference training set from which the resulting DP model is proven particularly suitable for accelerating calypso structure prediction. The scheme has been systematically benchmarked on binary magnesium-aluminium (Mg-Al) alloys and ternary lithium-lanthanum-hydrogen (Li-La-H) superhydrides, demonstrating its efficiency and reliability in DP model construction and calypso structure prediction. The proposed scheme represents a promising routine to perform the structure prediction of large or multicomponent systems.

关键词： Crystal structure Potential energy surfaces Machine learning

Efficient Privacy-Preserving ML for IoT: Cluster-Based Split Federated Learning Scheme for Non-IID Data

学校读者我要写书评

暂无评论

Efficient Privacy-Preserving ML for IoT: Cluster-Based Split...

Cyber Security in Networking Conference (CSNet)

作者： Mohamad Arafeh Mohamad Wazzeh Hakima Ould-Slimane Chamseddine Talhi Azzam Mourad Hadi Otrok Department of Software and IT engineering Ecole de Technologie Superieure (ETS) Montreal Canada Department of Mathematics and Computer Science Universite de Quebec a Trois-Rivieres (UQTR) Canada Division of Science New York University Abu Dhabi UAE Department of CSM Cyber Security Systems and Applied AI Research Center Lebanese American University Beirut Lebanon Department of EECS Center of Cyber-Physical Systems (C2PS) Khalifa University Abu Dhabi UAE

This paper proposes a scheme addressing the challenges of integrating privacy-preserving distributed machine learning in the Internet of Things (IoT) context while improving the efficiency of the learning process and accelerating the convergence of the model. Specifically, it addresses the presence of non-independent and identically distributed (Non-IID) data and device heterogeneity, which occurs during parallel training and are common characteristics of IoT environments. In this context, SplitFed is a privacy-preserving technique that combines Federated Learning (FL) and Split Learning (SL) to enable fast and privacy-preserving model training on IoT devices. Through parallelism, FL allows fast and parallel model training. Furthermore, SL enables weak devices to train models through model splitting. However, SplitFed suffers from the increasing convergence time due to the FL parallelism introducing Non-IID and stragglers issues. In this paper, we aim to address the aforementioned problems by proposing a clustering parallelization scheme. As such, similar clients are combined based on their weight divergence and considered a single FL unit addressing the Non-IID problem. Moreover, to improve time efficiency and overcome the stragglers' presence, we offer a second clustering stage that separates clients by efficiency and uses FedAsync for model aggregation.

关键词：

Optimized Security Mechanism for publicly Secret Key Sharing over Cloud using Blockchain

学校读者我要写书评

暂无评论

Optimized Security Mechanism for publicly Secret Key Sharing...

2024 International Symposium on Securing Next Generation Systems Using Future Artificial Intelligence Technologies, SNSFAIT 2024

作者： Byeon, Haewon Quraishi, Aadam Thuniki, Praveen Keshta, Ismail Soni, Mukesh Shabaz, Mohammad Department of AI and Software Inje University Gimhae50834 Korea Republic of Inje University Medical Big Data Research Center Gimhae50834 Korea Republic of Intervention Treatment Institute HoustonTX United States Sr Program Analyst Georgia CummingGA30040 United States Computer Science and Information Systems Department College of Applied Sciences Al Maarefa University Riyadh Saudi Arabia Dr. D. Y. Patii Vidyapeeth Pune Dr. D. Y. Patii School of Science & Technology Tathawade Pune India Model Institute of Engineering and Technology J&K Jammu India

A publicly verifiable key sharing mechanism based on threshold key sharing is provided to explore the security of users' private keys on the blockchain. Participating nodes check the key fragment after receiving it, effectively preventing it from being abused. The crucial sections of the nodes that participated in the critical splicing are made public during the critical recovery stage to prevent them from performing harmful things during the critical recovery stage. Add IDs to the nodes that participated in the crucial splicing during the key distribution stage;a dynamic threshold system is intended to track and update the status of malicious nodes in real time. When the node that possesses the crucial component fails, the owner of the critical component and the main node relocate a key element to the new participating nodes. To safeguard sensitive information. The experimental results show that this system has a key recovery rate of 80% and threshold qualities such as traceability, enforceability, and recoverability. © 2024 Copyright for this paper by its authors.

关键词： Blockchain Encryption Key sharing mechanism Privacy Secret key Security

End-to-end symmetry preserving inter-atomic potential energy model for finite and extended systems

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Zhang, Linfeng Han, Jiequn Wang, Han Saidi, Wissam A. Car, Roberto Weinan, E. Program in Applied and Computational Mathematics Princeton University United States Institute of Applied Physics and Computational Mathematics China CAEP Software Center for High Performance Numerical Simulation China Department of Mechanical Engineering and Materials Science University of Pittsburgh United States Department of Chemistry and Department of Physics Princeton University United States Princeton Institute for the Science and Technology of Materials Princeton University United States Department of Mathematics Princeton University United States Beijing Institute of Big Data Research China

Machine learning models are changing the paradigm of molecular modeling, which is a fundamental tool for material science, chemistry, and computational biology. Of particular interest is the inter-atomic potential energy surface (PES). Here we develop Deep Potential - Smooth Edition (DeepPot-SE), an end-to-end machine learning-based PES model, which is able to efficiently represent the PES of a wide variety of systems with the accuracy of ab initio quantum mechanics models. By construction, DeepPot-SE is extensive and continuously differentiable, scales linearly with system size, and preserves all the natural symmetries of the system. Further, we show that DeepPot-SE describes finite and extended systems including organic molecules, metals, semiconductors, and insulators with high fidelity. Copyright © 2018, The Authors. All rights reserved.

关键词： Potential energy

Task-Oriented Convex Bilevel Optimization with Latent Feasibility

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Liu, Risheng Ma, Long Yuan, Xiaoming Zeng, Shangzhi Zhang, Jin DUT-RU International School of Information Science & Engineering Dalian University of Technology Dalian116024 China The School of Software Technology Dalian University of Technology Dalian116024 China The Peng Cheng Laboratory Shenzhen518052 China The Department of Mathematics The University of Hong Kong Hong Kong The Department of Mathematics and Statistics University of Victoria Canada The Department of Mathematics SUSTech International Center for Mathematics Southern University of Science and Technology National Center for Applied Mathematics Shenzhen Shenzhen518055 China

This paper firstly proposes a convex bilevel optimization paradigm to formulate and optimize popular learning and vision problems in real-world scenarios. Different from conventional approaches, which directly design their iteration schemes based on given problem formulation, we introduce a task-oriented energy as our latent constraint which integrates richer task information. By explicitly re-characterizing the feasibility, we establish an efficient and flexible algorithmic framework to tackle convex models with both shrunken solution space and powerful auxiliary (based on domain knowledge and data distribution of the task). In theory, we present the convergence analysis of our latent feasibility re-characterization based numerical strategy. We also analyze the stability of the theoretical convergence under computational error perturbation. Extensive numerical experiments are conducted to verify our theoretical findings and evaluate the practical performance of our method on different applications. Copyright © 2019, The Authors. All rights reserved.

关键词： Convex optimization

Theory of thin-film-mediated exfoliation of van der Waals bonded layered materials

学校读者我要写书评

暂无评论

Physical Review Materials 2018年第9期2卷 094004-094004页

作者： Haoye Sun Eric W. Sirott James Mastandrea Hannah M. Gramling Yuzhi Zhou Max Poschmann Hayden K. Taylor Joel W. Ager D. C. Chrzan Department of Materials Science and Engineering University of California Berkeley California 94720 USA Materials Sciences Division Lawrence Berkeley National Laboratory Berkeley California 94720 USA Department of Mechanical Engineering University of California Berkeley California 94720 USA Laboratory of Computational Physics Huayuan Road 6 Beijing 100088 People's Republic of China Institute of Applied Physics and Computational Mathematics Fenghao East Road 2 Beijing 100094 People's Republic of China CAEP Software Center for High Performance Numerical Simulation Huayuan Road 6 Beijing 100088 People's Republic of China

Tape-mediated mechanical exfoliation of layered materials is known to require multiple exfoliation steps and to produce small, irregularly shaped monolayer samples. Recently, it has been demonstrated experimentally that deposition of a thin layer of Au on the layer to be exfoliated prior to application of the tape can yield large monolayer samples of transition metal dichalcogenides in one exfoliation step. Here the mechanism underlying these improvements is explored using atomic scale total energy calculations to study thin-film-assisted mechanical exfoliation of both MoS2 and graphene. This study focuses on the influence of epitaxially induced biaxial strain on the exfoliated layer and how this impacts monolayer selectivity during exfoliation. For graphite modeled with a reactive empirical bond order (REBO) plus a Lennard-Jones potential, tensile biaxial strain favors monolayer exfoliation, whereas compressive strain favors multilayer exfoliation. For graphite modeled with a more accurate registry-dependent potential and for MoS2 modeled with a REBO plus a Lennard-Jones potential, both compressive and tensile biaxial strains favor monolayer exfoliation. A simple model based on the interlayer potential is introduced to explain these observations, and a linear stability analysis of this model shows that the differences between the thin-film-assisted exfoliation of the considered materials arise from the stacking structure of the layered material and the details of the interlayer potential. The potential for using this simple model to screen suitable metal films for exfoliation of other 2D materials is discussed.

关键词： 2-dimensional systems Transition metal dichalcogenides

Optimization-Derived Learning with Essential Convergence Analysis of Training and Hyper-training

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Liu, Risheng Liu, Xuan Zeng, Shangzhi Zhang, Jin Zhang, Yixuan DUT-RU International School of Information Science and Engineering Dalian University of Technology Liaoning Dalian China Key Laboratory for Ubiquitous Network and Service Software of Liaoning Province Liaoning Dalian China Peng Cheng Laboratory Guangdong Shenzhen China Department of Mathematics and Statistics University of Victoria VictoriaBC Canada Department of Mathematics SUSTech International Center for Mathematics Southern University of Science and Technology Guangdong Shenzhen China National Center for Applied Mathematics Shenzhen Guangdong Shenzhen China

Recently, Optimization-Derived Learning (ODL) has attracted attention from learning and vision areas, which designs learning models from the perspective of optimization. However, previous ODL approaches regard the training and hyper-training procedures as two separated stages, meaning that the hyper-training variables have to be fixed during the training process, and thus it is also impossible to simultaneously obtain the convergence of training and hyper-training variables. In this work, we design a Generalized Krasnoselskii-Mann (GKM) scheme based on fixed-point iterations as our fundamental ODL module, which unifies existing ODL methods as special cases. Under the GKM scheme, a Bilevel Meta Optimization (BMO) algorithmic framework is constructed to solve the optimal training and hyper-training variables together. We rigorously prove the essential joint convergence of the fixed-point iteration for training and the process of optimizing hyper-parameters for hyper-training, both on the approximation quality, and on the stationary analysis. Experiments demonstrate the efficiency of BMO with competitive performance on sparse coding and real-world applications such as image deconvolution and rain streak removal. © 2022, CC BY-SA.

关键词： Iterative methods