arXiv subjects

Xin Wang

Publications and source records attributed to Xin Wang.

At least 1,405 records · Page 78Linked to original sources

arXiv2021

A preliminary study about gravitational wave radiation and cosmic heat death

We study the role of gravitational waves (GW) in the heat death of the universe. Due to the GW emission, in a very long period, dynamical systems in the universe suffer from persistent mechanical energy dissipation, evolving to a state of universal rest and death. With N-body simulations, we adopt a simple yet representative scheme to calculate the energy loss due to the GW emission. For current dark matter systems with mass $\sim10^{12}-10^{15} M_\odot$, we estimate their GW emission timescale as $\sim10^{19}-10^{25}$ years. This timescale is significantly larger than any baryon processes in the universe, but still $\sim10^{80}$ times shorter than that of the Hawking radiation. We stress that our analysis could be invalid due to many unknowns such as the dynamical chaos, the quadrupole momentum of halos, the angular momentum loss, the dynamic friction, the central black hole accretion, the dark matter decays or annihilations, the property of dark energy and the future evolution of the universe.

astro-ph.CO↗

arXiv2021

Infrastructure Assisted Constrained Connected Automated Vehicle Trajectory Optimization on Curved Roads: A Spatial Formulation on a Curvilinear Coordinate

Vehicle trajectory optimization is essential to ensure vehicles travel efficiently and safely. This paper presents an infrastructure assisted constrained connected automated vehicles (CAVs) trajectory optimization method on curved roads. This paper systematically formulates the problem based on a curvilinear coordinate which is flexible to model complex road geometries. Further, to deal with the spatial varying road obstacles, traffic regulations, and geometric characteristics, two-dimensional vehicle kinematics is given in a spatial formulation with exact road information provided by the infrastructure. Consequently, we applied a multi-objective model predictive control (MPC) approach to optimize the trajectories in a rolling horizon while satisfying the collision avoidances and vehicle kinematics constraints. To verify the efficiency of our method, a numerical simulation is conducted. As the results suggest, the proposed method can provide smooth vehicular trajectories, avoid road obstacles, and simultaneously follow traffic regulations, which is robust to road geometries and disturbances.

eess.SY↗

arXiv2021

When does reinforcement learning stand out in quantum control? A comparative study on state preparation

Reinforcement learning has been widely used in many problems, including quantum control of qubits. However, such problems can, at the same time, be solved by traditional, non-machine-learning methods, such as stochastic gradient descent and Krotov algorithms, and it remains unclear which one is most suitable when the control has specific constraints. In this work, we perform a comparative study on the efficacy of three reinforcement learning algorithms: tabular Q-learning, deep Q-learning, and policy gradient, as well as two non-machine-learning methods: stochastic gradient descent and Krotov algorithms, in the problem of preparing a desired quantum state. We found that overall, the deep Q-learning and policy gradient algorithms outperform others when the problem is discretized, e.g. allowing discrete values of control, and when the problem scales up. The reinforcement learning algorithms can also adaptively reduce the complexity of the control sequences, shortening the operation time and improving the fidelity. Our comparison provides insights into the suitability of reinforcement learning in quantum control problems.

quant-ph↗

arXiv2021

Learning Continuous-Time Dynamics by Stochastic Differential Networks

Learning continuous-time stochastic dynamics is a fundamental and essential problem in modeling sporadic time series, whose observations are irregular and sparse in both time and dimension. For a given system whose latent states and observed data are high-dimensional, it is generally impossible to derive a precise continuous-time stochastic process to describe the system behaviors. To solve the above problem, we apply Variational Bayesian method and propose a flexible continuous-time stochastic recurrent neural network named Variational Stochastic Differential Networks (VSDN), which embeds the complicated dynamics of the sporadic time series by neural Stochastic Differential Equations (SDE). VSDNs capture the stochastic dependency among latent states and observations by deep neural networks. We also incorporate two differential Evidence Lower Bounds to efficiently train the models. Through comprehensive experiments, we show that VSDNs outperform state-of-the-art continuous-time deep learning models and achieve remarkable performance on prediction and interpolation tasks for sporadic time series.

cs.LG↗

arXiv2021

Generalizable control for multiparameter quantum metrology

Quantum control can be employed in quantum metrology to improve the precision limit for the estimation of unknown parameters. The optimal control, however, typically depends on the actual values of the parameters and thus needs to be designed adaptively with the updated estimations of those parameters. Traditional methods, such as gradient ascent pulse engineering (GRAPE), need to be rerun for each new set of parameters encountered, making the optimization costly, especially when many parameters are involved. Here we study the generalizability of optimal control, namely, optimal controls that can be systematically updated across a range of parameters with minimal cost. In cases where control channels can completely reverse the shift in the Hamiltonian due to a change in parameters, we provide an analytical method which efficiently generates optimal controls for any parameter starting from an initial optimal control found by either GRAPE or reinforcement learning. When the control channels are restricted, the analytical scheme is invalid, but reinforcement learning still retains a level of generalizability, albeit in a narrower range. In cases where the shift in the Hamiltonian is impossible to decompose to available control channels, no generalizability is found for either the reinforcement learning or the analytical scheme. We argue that the generalization of reinforcement learning is through a mechanism similar to the analytical scheme. Our results provide insights into when and how the optimal control in multiparameter quantum metrology can be generalized, thereby facilitating efficient implementation of optimal quantum estimation of multiple parameters, particularly for an ensemble of systems with ranges of parameters.

quant-ph↗

arXiv2021

Generalizable control for quantum parameter estimation through reinforcement learning

Measurement and estimation of parameters are essential for science and engineering, where one of the main quests is to find systematic schemes that can achieve high precision. While conventional schemes for quantum parameter estimation focus on the optimization of the probe states and measurements, it has been recently realized that control during the evolution can significantly improve the precision. The identification of optimal controls, however, is often computationally demanding, as typically the optimal controls depend on the value of the parameter which then needs to be re-calculated after the update of the estimation in each iteration. Here we show that reinforcement learning provides an efficient way to identify the controls that can be employed to improve the precision. We also demonstrate that reinforcement learning is highly generalizable, namely the neural network trained under one particular value of the parameter can work for different values within a broad range. These desired features make reinforcement learning an efficient alternative to conventional optimal quantum control methods.

quant-ph↗

arXiv2021

Charge noise suppression in capacitively coupled singlet-triplet spin qubits under magnetic field

Charge noise is the main hurdle preventing high-fidelity operation, in particular that of two-qubit gates, of semiconductor-quantum-dot-based spin qubits. While certain sweet spots where charge noise is substantially suppressed have been demonstrated in several types of spin qubits, the existence of one for coupled singlet-triplet qubits is unclear. We theoretically demonstrate, using full configuration-interaction calculations, that a range of nearly sweet spots appear in the coupled singlet-triplet qubit system when a strong enough magnetic field is applied externally. We further demonstrate that ramping to and from the judiciously chosen nearly sweet spot using sequences based on the shortcut to adiabaticity offers maximal gate fidelities under charge noise and phonon-induced decoherence. These results should facilitate realization of high-fidelity two-qubit gates in singlet-triplet qubit systems.

cond-mat.mes-hall↗

arXiv2021

Noise-Assisted Quantum Autoencoder

Quantum autoencoder is an efficient variational quantum algorithm for quantum data compression. However, previous quantum autoencoders fail to compress and recover high-rank mixed states. In this work, we discuss the fundamental properties and limitations of the standard quantum autoencoder model in more depth, and provide an information-theoretic solution to its recovering fidelity. Based on this understanding, we present a noise-assisted quantum autoencoder algorithm to go beyond the limitations, our model can achieve high recovering fidelity for general input states. Appropriate noise channels are used to make the input mixedness and output mixedness consistent, the noise setup is determined by measurement results of the trash system. Compared with the original quantum autoencoder model, the measurement information is fully used in our algorithm. In addition to the circuit model, we design a (noise-assisted) adiabatic model of quantum autoencoder that can be implemented on quantum annealers. We verified the validity of our methods through compressing the thermal states of transverse field Ising model and Werner states. For pure state ensemble compression, we also introduce a projected quantum autoencoder algorithm.

quant-ph↗

arXiv2021

Large Amplitude Switchback Turbulence: Possible Magnetic Velocity Alignment Structures

Switchbacks are widely acknowledged phenomena observed by the Parker Solar Probe and appear to occur in patches. Previous studies focused on the fluctuations at the magnetic reversals. However, the nature of the fluctuations inside the switchbacks remains unknown. Here we utilize the magnetic field data and plasma data measured by the Parker Solar Probe in the first four encounters. We investigate the fluctuations in the switchback intervals of 100 s with BR>0 at every instant and compare them to the fluctuations in the nonswitchback intervals of 100 s with theta_RB>160o at every instant. We calculate normalized cross-helicity sigma_c, normalized residual energy sigma_r, correlation coefficient C_vb between dvA and dv, Alfven ratio rA, and the amplitude of magnetic and kinetic fluctuations. We find that the switchback intervals exhibit a distribution of sigma_c similar with the nonswitchback intervals. However, the rA of switchback intervals is around 0.35, while the nonswitchback intervals have rA around 0.65, indicating the fluctuations in the switchbacks are more magnetically dominated. We also find that the distribution pattern of pixel average amplitude of both dvA and dv of switchback intervals in the C_vb-sigma_r plane show a vertical stripe feature at C_vb>0.8, illustrating the possible magnetically dominant magnetic-velocity alignment structure. These results will help us to understand the nature and the formation of the switchback turbulence.

astro-ph.SR↗

arXiv2021

Generic detection-based error-mitigation using quantum autoencoders

Efficient error-mitigation techniques demanding minimal resources is key to quantum information processing. We propose a generic protocol to mitigate quantum errors using detection-based quantum autoencoders. In our protocol, the quantum data are compressed into a latent subspace while leaving errors outside, the latter of which is then removed by a measurement and post-selection. Compared to previously developed methods, our protocol on the one hand requires no extra qubits, and on the other hand has a near-optimal denoising power, in which under reasonable requirements all errors detected outside of the latent subspace can be removed, while those inside the subspace cannot be removed by any means. Our detection-based quantum autoencoders are therefore particularly useful for near-term quantum devices in which controllable qubits are limited while noise reduction is important.

quant-ph↗

arXiv2021

Relational Subsets Knowledge Distillation for Long-tailed Retinal Diseases Recognition

In the real world, medical datasets often exhibit a long-tailed data distribution (i.e., a few classes occupy most of the data, while most classes have rarely few samples), which results in a challenging imbalance learning scenario. For example, there are estimated more than 40 different kinds of retinal diseases with variable morbidity, however with more than 30+ conditions are very rare from the global patient cohorts, which results in a typical long-tailed learning problem for deep learning-based screening models. In this study, we propose class subset learning by dividing the long-tailed data into multiple class subsets according to prior knowledge, such as regions and phenotype information. It enforces the model to focus on learning the subset-specific knowledge. More specifically, there are some relational classes that reside in the fixed retinal regions, or some common pathological features are observed in both the majority and minority conditions. With those subsets learnt teacher models, then we are able to distill the multiple teacher models into a unified model with weighted knowledge distillation loss. The proposed framework proved to be effective for the long-tailed retinal diseases recognition task. The experimental results on two different datasets demonstrate that our method is flexible and can be easily plugged into many other state-of-the-art techniques with significant improvements.

cs.CV↗

arXiv2021

Understanding the Benefit of Being Patient in Payment Channel Networks

Scaling blockchain efficiency is crucial to its widespread usage in which the payment channel is one of the most prominent approaches. With payment channels and the network they construct, two users can move some transactions off-chain in a predetermined duration to avoid expensive and time-consuming on-chain settlements. Existing work is devoted to designing high-throughput payment channel networks (PCNs) or efficient PCN routing policies to reduce the fee. In this paper, we investigate the PCN routing from a different perspective by answering whether the routing fee of transactions can be saved by being a bit more patient. The key idea is to reorder the processing sequence of atomic transactions, other than to handle each of them separately and immediately. We present two mechanisms, one is periodic transaction processing assisted by a PCN broker and the other is purely strategic waiting. In the former, all the incoming transactions in a short time interval are processed collectively. We formulate an optimization model to minimize their total routing fee and derive the optimal permutation of processing transactions as well as the routing policy for each of them. A Shapley value based scheme is presented to redistribute the benefit of reordering among the transactions efficiently and fairly. In the latter, we model the waiting time of a strategic transaction on a single payment channel as the first passage time problem in queueing theory when the transaction value is higher than the edge capacity upon its arrival. By capturing the capacity dynamics, we are able to calculate the recursive expression of waiting for time distribution that is useful to gauge a user's cost of patience. Experimental results manifest that our cost redistribution mechanism can effectively save routing fees for all the transactions, and the waiting time distribution coincides with the model well.

cs.GT↗

arXiv2021

Spontaneous imbibition in porous media: from pore scale to Darcy scale

Spontaneous imbibition has been receiving much attention due to its significance in many subsurface and industrial applications. Unveiling pore-scale wetting dynamics, and particularly its upscaling to the Darcy scale are still unresolved. In this work, we conduct image-based pore-network modeling of cocurrent spontaneous imbibition and the corresponding quasi-static imbibition, in homogeneous sintered glass beads as well as heterogeneous Estaillades. A wide range of viscosity ratios and wettability conditions are taken into account. Based on our pore-scale results, we show the influence of pore-scale heterogeneity on imbibition dynamics and nonwetting entrapment. We elucidate different pore-filling mechanisms in imbibition, which helps us understand wetting dynamics. Most importantly, we develop a non-equilibrium model for relative permeability of the wetting phase, which adequately incorporates wetting dynamics. This is crucial to the final goal of developing a two-phase imbibition model with measurable material properties such as capillary pressure and relative permeability. Finally, we propose some future work on both numerical and experimental verifications of the developed non-equilibrium permeability model.

cond-mat.soft↗

arXiv2021

Multi-Scale Context Aggregation Network with Attention-Guided for Crowd Counting

Crowd counting aims to predict the number of people and generate the density map in the image. There are many challenges, including varying head scales, the diversity of crowd distribution across images and cluttered backgrounds. In this paper, we propose a multi-scale context aggregation network (MSCANet) based on single-column encoder-decoder architecture for crowd counting, which consists of an encoder based on a dense context-aware module (DCAM) and a hierarchical attention-guided decoder. To handle the issue of scale variation, we construct the DCAM to aggregate multi-scale contextual information by densely connecting the dilated convolution with varying receptive fields. The proposed DCAM can capture rich contextual information of crowd areas due to its long-range receptive fields and dense scale sampling. Moreover, to suppress the background noise and generate a high-quality density map, we adopt a hierarchical attention-guided mechanism in the decoder. This helps to integrate more useful spatial information from shallow feature maps of the encoder by introducing multiple supervision based on semantic attention module (SAM). Extensive experiments demonstrate that the proposed approach achieves better performance than other similar state-of-the-art methods on three challenging benchmark datasets for crowd counting. The code is available at https://github.com/KingMV/MSCANet

cs.CV↗

arXiv2021

Online Learning of a Probabilistic and Adaptive Scene Representation

Constructing and maintaining a consistent scene model on-the-fly is the core task for online spatial perception, interpretation, and action. In this paper, we represent the scene with a Bayesian nonparametric mixture model, seamlessly describing per-point occupancy status with a continuous probability density function. Instead of following the conventional data fusion paradigm, we address the problem of online learning the process how sequential point cloud data are generated from the scene geometry. An incremental and parallel inference is performed to update the parameter space in real-time. We experimentally show that the proposed representation achieves state-of-the-art accuracy with promising efficiency. The consistent probabilistic formulation assures a generative model that is adaptive to different sensor characteristics, and the model complexity can be dynamically adjusted on-the-fly according to different data scales.

cs.CV↗

arXiv2021

One Network Fits All? Modular versus Monolithic Task Formulations in Neural Networks

Can deep learning solve multiple tasks simultaneously, even when they are unrelated and very different? We investigate how the representations of the underlying tasks affect the ability of a single neural network to learn them jointly. We present theoretical and empirical findings that a single neural network is capable of simultaneously learning multiple tasks from a combined data set, for a variety of methods for representing tasks -- for example, when the distinct tasks are encoded by well-separated clusters or decision trees over certain task-code attributes. More concretely, we present a novel analysis that shows that families of simple programming-like constructs for the codes encoding the tasks are learnable by two-layer neural networks with standard training. We study more generally how the complexity of learning such combined tasks grows with the complexity of the task codes; we find that combining many tasks may incur a sample complexity penalty, even though the individual tasks are easy to learn. We provide empirical support for the usefulness of the learning bounds by training networks on clusters, decision trees, and SQL-style aggregation.

cs.LG↗

arXiv2021

A Survey on Curriculum Learning

Curriculum learning (CL) is a training strategy that trains a machine learning model from easier data to harder data, which imitates the meaningful learning order in human curricula. As an easy-to-use plug-in, the CL strategy has demonstrated its power in improving the generalization capacity and convergence rate of various models in a wide range of scenarios such as computer vision and natural language processing etc. In this survey article, we comprehensively review CL from various aspects including motivations, definitions, theories, and applications. We discuss works on curriculum learning within a general CL framework, elaborating on how to design a manually predefined curriculum or an automatic curriculum. In particular, we summarize existing CL designs based on the general framework of Difficulty Measurer+Training Scheduler and further categorize the methodologies for automatic CL into four groups, i.e., Self-paced Learning, Transfer Teacher, RL Teacher, and Other Automatic CL. We also analyze principles to select different CL designs that may benefit practical applications. Finally, we present our insights on the relationships connecting CL and other machine learning concepts including transfer learning, meta-learning, continual learning and active learning, etc., then point out challenges in CL as well as potential future research directions deserving further investigations.

cs.LG↗

arXiv2021

Measurement Error Mitigation via Truncated Neumann Series

Measurements on near-term quantum processors are inevitably subject to hardware imperfections that lead to readout errors. Mitigation of such unavoidable errors is crucial to better explore and extend the power of near-term quantum hardware. In this work, we propose a method to mitigate measurement errors in computing quantum expectation values using the truncated Neumann series. The essential idea is to cancel the errors by combining various noisy expectation values generated by sequential measurements determined by terms in the truncated series. We numerically test this method and find that the computation accuracy is substantially improved. Our method possesses several advantages: it does not assume any noise structure, it does not require the calibration procedure to learn the noise matrix a prior, and most importantly, the incurred error mitigation overhead is independent of system size, as long as the noise resistance of the measurement device is moderate. All these advantages empower our method as a practical measurement error mitigation method for near-term quantum devices.

quant-ph↗