检索结果-内蒙古大学图书馆

arXiv 2019年

作者： Chattopadhyay, Ashesh Hassanzadeh, Pedram Palem, Krishna Subramanian, Devika Department of Mechanical Engineering Rice University Department of Earth Environmental and Planetary Sciences Rice University Departments of Computer Science and Electrical and Computer Engineering Rice University

In this paper, the performance of three deep learning methods for predicting short-term evolution and reproducing the long-term statistics of a multi-scale spatio-temporal Lorenz 96 system is examined. The methods are: echo state network (a type of reservoir computing, RC-ESN), deep feed-forward artificial neural network (ANN), and recurrent neural network with long short-term memory (RNN- LSTM). This Lorenz system has three tiers of nonlinearly interacting variables representing slow/large- scale (X), intermediate (Y ), and fast/small-scale (Z) processes. For training or testing, only X is available;Y and Z are never known/used. It is shown that RC-ESN substantially outperforms ANN and RNN-LSTM for short-term prediction, e.g., accurately forecasting the chaotic trajectories for hundreds of numerical solver's time steps, equivalent to several Lyapunov timescales. RNN-LSTM and ANN show some prediction skills as well;RNN-LSTM bests ANN. Furthermore, even after losing the trajectory, data predicted by RC-ESN and RNN-LSTM have probability density functions (PDFs) that closely match the true PDF, even at the tails. PDF of the ANN data deviates from the true PDF. Implications, caveats, and applications to data-driven and inexact, data-assisted surrogate modeling of complex dynamical systems such as weather/climate are discussed. Copyright © 2019, The Authors. All rights reserved.

关键词： Probability density function

Fault-tolerant operation of a quantum error-correction code

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Egan, Laird Debroy, Dripto M. Noel, Crystal Risinger, Andrew Zhu, Daiwei Biswas, Debopriyo Newman, Michael Li, Muyuan Brown, Kenneth R. Cetina, Marko Monroe, Christopher Joint Quantum Institute Center for Quantum Information and Computer Science Departments of Physics and Electrical and Computer Engineering University of Maryland College ParkMD20742 United States Department of Physics Duke University DurhamNC27708 United States Department of Electrical and Computer Engineering Duke University DurhamNC27708 United States Department of Chemistry Duke University DurhamNC27708 United States Schools of Chemistry and Biochemistry and Computational Science and Engineering Georgia Institute of Technology AtlantaGA30332 United States Google Research VeniceCA90291 United States

Quantum error correction protects fragile quantum information by encoding it in a larger quantum system whose extra degrees of freedom enable the detection and correction of errors. An encoded logical qubit thus carries increased complexity compared to a bare physical qubit. Fault-tolerant protocols contain the spread of errors and are essential for realizing error suppression with an error-corrected logical qubit. Here we experimentally demonstrate fault-tolerant preparation, rotation, error syndrome extraction, and measurement on a logical qubit encoded in the 9-qubit Bacon-Shor code. For the logical qubit, we measure an average fault-tolerant preparation and measurement error of 0.6% and a transversal Clifford gate with an error of 0.3% after error correction. The result is an encoded logical qubit whose logical fidelity exceeds the fidelity of the entangling operations used to create it. We compare these operations with non-fault-tolerant protocols capable of generating arbitrary logical states, and observe the expected increase in error. We directly measure the four Bacon-Shor stabilizer generators and are able to detect single qubit Pauli errors. These results show that fault-tolerant quantum systems are currently capable of logical primitives with error rates lower than their constituent parts. With the future addition of intermediate measurements, the full power of scalable quantum error-correction can be achieved. Copyright © 2020, The Authors. All rights reserved.

关键词： Quantum optics

Corrections to "iPhantom: A Framework for Automated Creation of Individualized Computational Phantoms and its Application to CT Organ Dosimetry"

学校读者我要写书评

暂无评论

IEEE journal of biomedical and health informatics 2022年第1期26卷 478页

作者： Wanyi Fu Shobhit Sharma Ehsan Abadi Alexandros-Stavros Iliopoulos Qi Wang Joseph Y Lo Xiaobai Sun William P Segars Ehsan Samei Department of Electrical and Computer Engineering and Carl E. Ravin Advanced Imaging Laboratories Duke University Durham NC USA Department of Physics and Carl E. Ravin Advanced Imaging Laboratories Duke University Durham NC USA Department of Computer Science Duke University Durham NC USA Computer Science and Artificial Intelligence Laboratory Massachusetts Institute of Technology Cambridge MA USA Department of Radiology the Fourth Clinical Hospital of Hebei Medical University Heibei China Departments of Electrical and Computer Engineering Biomedical Engineering Medical Physics Graduate Program and Carl E. Ravin Advanced Imaging Laboratories Duke University Durham NC USA Medical Physics Graduate Program and Radiology and the Carl E. Ravin Advanced Imaging Laboratories Departments of Biomedical Engineering Duke University Durham NC USA Carl E. Ravin Advanced Imaging Laboratories Medical Physics Graduate Program Department of Radiology Electrical and Computer Engineering Biomedical Engineering and Physics Duke University Durham NC USA

In [1], the dose estimation accuracy using the alternative baseline method under modulated tube current was not correctly calculated due to an unintentional simulation error.

关键词： Phantoms Image segmentation Computed tomography Image registration

Alibaba cloud quantum development platform: Surface code simulations with crosstalk

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Huang, Cupjin Ni, Xiaotong Zhang, Fang Newman, Michael Ding, Dawei Gao, Xun Wang, Tenghui Zhao, Hui-Hai Wu, Feng Zhang, Gengyan Deng, Chunqing Ku, Hsiang-Sheng Chen, Jianxin Shi, Yaoyun Alibaba Quantum Laboratory Alibaba Group USA BellevueWA98004 United States Alibaba Quantum Laboratory Alibaba Group Hangzhou Zhejiang311121 Department of Electrical Engineering and Computer Science University of Michigan Ann ArborMI48109 United States Departments of Physics and Electrical and Computer Engineering Duke University DurhamNC27708 United States Alibaba Quantum Laboratory Alibaba Group Beijing100102

We report, in a sequence of notes, our work on the Alibaba Cloud Quantum Development Platform (AC-QDP). AC-QDP provides a set of tools for aiding the development of both quantum computing algorithms and quantum processors, and is powered by a large-scale classical simulator deployed on Alibaba Cloud. In this note, we simulate a distance-3 logical qubit encoded in the 17-qubit surface code using experimental noise parameters for transmon qubits in a planar circuit QED architecture. Our simulation features crosstalk induced by ZZ-interactions. We show that at current noise levels, crosstalk contributes significantly to the dephasing of the logical qubit. This results in a total phase-flip probability of ∼ 0:6%, about 60% higher than expected without considering crosstalk. Our results quantify the significant damage caused by simple, correctable crosstalk, emphasizing the need to mitigate even its second-order effects in order to realize beneficial error-correction. Copyright © 2020, The Authors. All rights reserved.

关键词： Crosstalk

Floquet vortex states induced by light carrying the orbital angular momentum

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Kim, Hwanmun Dehghani, Hossein Ahmadabadi, Iman Martin, Ivar Hafezi, Mohammad Joint Quantum Institute NIST and University of Maryland College ParkMD20742 United States Department of Physics University of Maryland College ParkMD20742 United States Departments of Electrical and Computer Engineering Institute for Research in Electronics and Applied Physics University of Maryland College ParkMD20742 United States Materials Science Division Argonne National Laboratory ArgonneIL60439 United States

We propose a scheme to create an electronic Floquet vortex state by irradiating a two-dimensional semiconductor with the laser light carrying non-zero orbital angular momentum. We analytically and numerically study the properties of the Floquet vortex states, with the methods analogous to the ones previously applied to the analysis of superconducting vortex states. We show that such Floquet vortex states are similar to the superconducting vortex states, and they exhibit a wide range of tunability. To illustrate the potential utility of such tunability, we show how such states could be used for quantum state engineering. Copyright © 2021, The Authors. All rights reserved.

关键词： Angular momentum

H2NF-Net for Brain Tumor Segmentation using Multimodal MR Imaging: 2nd Place Solution to BraTS Challenge 2020 Segmentation Task

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Jia, Haozhe Cai, Weidong Huang, Heng Xia, Yong Research & Development Institute Northwestern Polytechnical University in Shenzhen Shenzhen518057 China National Engineering Laboratory for Integrated Aero-Space-Ground-Ocean Big Data Application Technology School of Computer Science and Engineering Northwestern Polytechnical University Xi’an710072 China School of Computer Science University of Sydney SydneyNSW2006 Australia Department of Electrical and Computer Engineering University of Pittsburgh PittsburghPA15261 United States JD Finance America Corporation CaliforniaCA94043 United States

In this paper, we propose a Hybrid High-resolution and Non-local Feature Network (H2NF-Net) to segment brain tumor in multimodal MR images. Our H2NF-Net uses the single and cascaded HNF-Nets to segment different brain tumor sub-regions and combines the predictions together as the final segmentation. We trained and evaluated our model on the Multimodal Brain Tumor Segmentation Challenge (BraTS) 2020 dataset. The results on the test set show that the combination of the single and cascaded models achieved average Dice scores of 0.78751, 0.91290, and 0.85461, as well as Hausdorff distances (95%) of 26.57525, 4.18426, and 4.97162 for the enhancing tumor, whole tumor, and tumor core, respectively. Our method won the second place in the BraTS 2020 challenge segmentation task out of nearly 80 participants. © 2020, CC BY.

关键词： Tumors

Multitask Deep Learning for Accurate Risk Stratification and Prediction of Next Steps for Coronary Ct Angiography Patients

学校读者我要写书评

暂无评论

SSRN

SSRN 2023年

作者： Lu, Juan Bennamoun, Mohammed Stewart, Jonathon Eshraghian, Jason K. Liu, Yanbin Chow, Benjamin Sanfilippo, Frank M. Dwivedi, Girish Medical School The University of Western Australia 35 Stirling Highway NedlandsWA6009 Australia School of Computer Science The University of Western Australia 35 Stirling Highway NedlandsWA6009 Australia Harry Perkins Institute of Medical Research 5 Robin Warren Dr MurdochWA6150 Australia Department of Electrical and Computer Engineering University of California 1156 High Street Santa CruzCA95064 United States School of Computing Australian National University Australian Capital Territory Canberra2601 Australia Division of Cardiology Departments of Medicine University of Ottawa Heart Institute 40 Ruskin Street OttawaONK1Y 4W7 Canada School of Population and Global Health The University of Western Australia 35 Stirling Highway NedlandsWA6009 Australia Department of Advanced Clinical and Translational Cardiovascular Imaging

Diagnostic investigation has an important role in risk stratification and clinical decision making of patients with suspected and documented coronary artery disease (CAD). However, the majority of existing tools are primarily focused on the selection of gatekeeper tests, whereas only a handful of systems contain information regarding the downstream testing or treatment. We propose a multi-task deep learning model to support risk stratification and down-stream test selection for patients undergoing coronary computed tomography angiography (CCTA). The analysis included 14,021 patients who underwent CCTA between 2006 and 2017. Our novel multitask deep learning framework (TabPerceiver) extends the state-of-the-art Perceiver model to deal with real-world patient information and their CCTA report data. Despite significant progress using deep learning models for tabular data, they rarely outperform gradient boosting decision trees (GBDT). Our multi-task TabPerceiver model achieved an area under the receiver operating characteristic curve (AUC) of 0.752 (low), 0.629 (medium), and 0.765 (high) in CAD risk stratification, and 0.730 in predicting downstream tests, which has outperformed the benchmark GBDT (low: 0.733, medium: 0.610, high: 0.719, downstream test: 0.741).To the best of our knowledge, our multi-task TabPerceiver model is the first cross-attention based deep learning model processing tabular data, it accurately estimates the patients’ risk of CAD and provides recommended downstream tests based on prior CCTA information. In clinical practice, the utilization of such an approach could bring a paradigm shift in risk stratification and downstream management. Our study also shows that neural networks appear to benefit more readily from multi-task learning than tree-based models. This could offset the shortcomings of using single-task deep learning approach when working with tabular data. © 2023, The Authors. All rights reserved.

关键词： Decision trees

Problems with classification, hypothesis testing, and estimator convergence in the analysis of degree distributions in networks

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Van Der Hoorn, Pim Voitalov, Ivan van der Hofstad, Remco Krioukov, Dmitri Department of Mathematics and Computer Science Eindhoven University of Technology Eindhoven Netherlands Department of Physics Northeastern University BostonMA02115 United States Network Science Institute Northeastern University BostonMA02115 United States Departments of Mathematics and Electrical and Computer Engineering Northeastern University BostonMA02115 United States

Model adaptation for inverse problems in imaging

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Gilton, Davis Ongie, Gregory Willett, Rebecca Department of Electrical and Computer Engineering University of Wisconsin-Madison 1415 Engineering Dr MadisonWI53706 United States Department of Mathematical and Statistical Sciences Marquette University 1250 W Wisconsin Ave MilwaukeeWI53233 United States Departments of Computer Science and Statistics University of Chicago 5747 S Ellis Ave ChicagoIL60637 United States

Deep neural networks have been applied successfully to a wide variety of inverse problems arising in computational imaging. These networks are typically trained using a forward model that describes the measurement process to be inverted, which is often incorporated directly into the network itself. However, these approaches are sensitive to changes in the forward model: if at test time the forward model varies (even slightly) from the one the network was trained for, the reconstruction performance can degrade substantially. Given a network trained to solve an initial inverse problem with a known forward model, we propose two novel procedures that adapt the network to a change in the forward model, even without full knowledge of the change. Our approaches do not require access to more labeled data (i.e., ground truth images). We show these simple model adaptation approaches achieve empirical success in a variety of inverse problems, including deblurring, super-resolution, and undersampled image reconstruction in magnetic resonance imaging. © 2020, CC BY.

关键词： Differential equations

electrically powered locomotion of dual-nature colloidhedgehog and colloid-umbilic topological and elastic dipoles in liquid crystals

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Senyuk, Bohdan Adufu, Richmond E. Smalyukh, Ivan I. Department of Physics University of Colorado BoulderCO80309 United States Department of Electrical Computer and Energy Engineering University of Colorado BoulderCO80309 United States Soft Materials Research Center Materials Science and Engineering Program University of Colorado BoulderCO80309 United States Chemical Physics Program Departments of Chemistry and Physics University of Colorado BoulderCO80309 United States Renewable and Sustainable Energy Institute National Renewable Energy Laboratory University of Colorado BoulderCO80309 United States

Colloidal particles in liquid crystals tend to induce topological defects and distortions of the molecular alignment within the surrounding anisotropic host medium, which results in elasticity-mediated interactions not accessible to their counterparts within isotropic fluid hosts. Such particle-induced coronae of perturbed nematic order are highly responsive to external electric fields, even when the uniformly aligned host medium away from particles exhibits no response to fields below the realignment threshold. Here we harness the non-reciprocal nature of these facile electric responses to demonstrate colloidal locomotion. Oscillations of electric field prompt repetitive deformations of the corona of dipolar elastic distortions around the colloidal inclusions, which at appropriately designed electric driving synchronize the displacement directions. We observe the colloid-hedgehog dipole accompanied by an umbilical defect in the tilt directionality field (c-field), along with the texture of elastic distortions that evolves with changing applied voltage. The temporal out-of-equilibrium evolution of the director and c-field distortions around particles upon turning voltage on and off is not invariant upon reversal of time, prompting lateral translations and interactions that markedly differ from what is accessible to these colloids under equilibrium conditions. Our findings may lead to both technological and fundamental science applications of nematic colloids as both model reconfigurable colloidal systems and as mesostructured materials with pre-designed temporal evolution of structure and composition. © 2021, CC BY-NC-ND.

关键词： Liquid crystals