arXiv subjects

Bin Wang

Publications and source records attributed to Bin Wang.

At least 559 records · Page 31Linked to original sources

arXiv2022

Dosimetric Evaluation of a New Rotating Gamma System for Stereotactic Radiosurgery

Purpose: A novel rotating gamma stereotactic radiosurgery (SRS) system (Galaxy RTi) with real-time image guidance technology has been developed for high-precision SRS and frameless fractionated stereotactic radiotherapy (SRT). This work investigated the dosimetric quality of Galaxy by comparing both the machine treatment parameters and plan dosimetry parameters with those of the widely used Leksell Gamma Knife (LGK) systems for SRS. Methods: The Galaxy RTi system uses 30 cobalt-60 sources on a rotating gantry to deliver non-coplanar, non-overlapping arcs simultaneously while the LGK 4C uses 201 static cobalt-60 sources to deliver noncoplanar beams. Ten brain cancer patients were unarchived from our clinical database, which were previously treated on the LGK 4C. The lesion volume for these cases varied from 0.1 cm3 to 15.4 cm3. Galaxy plans were generated using the Prowess TPS (Prowess, Concord, CA) with the same dose constraints and optimization parameters. Treatment quality metrics such as target coverage (%volume receiving the prescription dose), conformity index (CI), cone size, shots number, beam-on time were compared together with DVH curves and dose distributions. Results: Superior treatment plans were generated for the Galaxy system that met our clinical acceptance criteria. For the 10 patients investigated, the mean CI and dose coverage for Galaxy was 1.77 and 99.24 compared to 1.94 and 99.19 for LGK, respectively. The beam-on time for Galaxy was 17.42 minutes compared to 21.34 minutes for LGK (both assuming dose rates at the initial installation). The dose fall-off is much faster for Galaxy, compared with LGK. Conclusion: The Galaxy RTi system can provide dose distributions with similar quality to that of LGK with less beam-on time and faster dose fall-off. The system is also capable of real-time image guidance at treatment position to ensure accurate dose delivery for SRS.

physics.med-ph↗

arXiv2022

Decomposed Mutual Information Optimization for Generalized Context in Meta-Reinforcement Learning

Adapting to the changes in transition dynamics is essential in robotic applications. By learning a conditional policy with a compact context, context-aware meta-reinforcement learning provides a flexible way to adjust behavior according to dynamics changes. However, in real-world applications, the agent may encounter complex dynamics changes. Multiple confounders can influence the transition dynamics, making it challenging to infer accurate context for decision-making. This paper addresses such a challenge by Decomposed Mutual INformation Optimization (DOMINO) for context learning, which explicitly learns a disentangled context to maximize the mutual information between the context and historical trajectories, while minimizing the state transition prediction error. Our theoretical analysis shows that DOMINO can overcome the underestimation of the mutual information caused by multi-confounded challenges via learning disentangled context and reduce the demand for the number of samples collected in various environments. Extensive experiments show that the context learned by DOMINO benefits both model-based and model-free reinforcement learning algorithms for dynamics generalization in terms of sample efficiency and performance in unseen environments.

cs.LG↗

arXiv2022

Critical Dynamics in Holographic First-Order Phase Transition

We study the critical phenomena of the dynamical transition from a metastable state to a stable state in the model of first-order phase transition via two different triggering mechanisms. Three universal stages during the fully nonlinear evolution are extracted. On the one side, by perturbing the scalar source, an isolated seed nucleus is injected into an initial homogeneous state in the supercooled region. For critical parameters of the seed nucleus, the real-time dynamics reveal that the system will converge to a critically unstable state. For supercritical parameters, the system exhibits a phase separation, while for subcritical parameters falls back to homogeneous. The shape independence of the seed nucleus is also investigated, which implies that the critical phenomena are universal. On the other side, we propose a novel mechanism to render the critical phenomena via a collision of two gravitational shock waves on the dual geometries. Specifying a collision velocity, the initial system can be also attracted to a critically unstable state. Aside from these dynamical constructions, we also quantitatively analyze the critical nucleus preventing the system from reaching the final phase separation. We find the depth of the critical nucleus increases almost linearly with the temperature, which implies that the hotter the supercooled state is, the harder for it to trigger phase separation.

hep-th↗

arXiv2022

Dynamical descalarization in Einstein-Maxwell-scalar theory

For an asymptotically flat hairy black hole in the Einstein-Maxwell-scalar (EMS) theory, we study the possibility of shedding off its scalar hair via nonlinear scalar perturbation fully interacting with the background spacetime. We examine the effect of the perturbation strength on the descalarization. The results show that the effective charge to mass ratio of the black hole plays the key role in the dynamical descalarization. The descalarization at the threshold is continuous. This indicates a second order phase transition.

gr-qc↗

arXiv2022

HiSTGNN: Hierarchical Spatio-temporal Graph Neural Networks for Weather Forecasting

Weather Forecasting is an attractive challengeable task due to its influence on human life and complexity in atmospheric motion. Supported by massive historical observed time series data, the task is suitable for data-driven approaches, especially deep neural networks. Recently, the Graph Neural Networks (GNNs) based methods have achieved excellent performance for spatio-temporal forecasting. However, the canonical GNNs-based methods only individually model the local graph of meteorological variables per station or the global graph of whole stations, lacking information interaction between meteorological variables in different stations. In this paper, we propose a novel Hierarchical Spatio-Temporal Graph Neural Network (HiSTGNN) to model cross-regional spatio-temporal correlations among meteorological variables in multiple stations. An adaptive graph learning layer and spatial graph convolution are employed to construct self-learning graph and study hidden dependency among nodes of variable-level and station-level graph. For capturing temporal pattern, the dilated inception as the backbone of gate temporal convolution is designed to model long and various meteorological trends. Moreover, a dynamic interaction learning is proposed to build bidirectional information passing in hierarchical graph. Experimental results on three real-world meteorological datasets demonstrate the superior performance of HiSTGNN beyond 7 baselines and it reduces the errors by 4.2% to 11.6% especially compared to state-of-the-art weather forecasting method.

cs.LG↗

arXiv2022

Quantum frequency conversion and single-photon detection with lithium niobate nanophotonic chips

In the past few years, the lithium niobate on insulator (LNOI) platform has revolutionized lithium niobate materials, and a series of quantum photonic chips based on LNOI have shown unprecedented performances. Quantum frequency conversion (QFC) photonic chips, which enable quantum state preservation during frequency tuning, are crucial in quantum technology. In this work, we demonstrate a low-noise QFC process on an LNOI nanophotonic platform designed to connect telecom and near-visible bands with sum-frequency generation by long-wavelength pumping. An internal conversion efficiency of 73% and an on-chip noise count rate of 900 counts per second (cps) are achieved. Moreover, the on-chip preservation of quantum statistical properties is verified, showing that the QFC chip is promising for extensive applications of LNOI integrated circuits in quantum information. Based on the QFC chip, we construct an upconversion single-photon detector with the sum-frequency output spectrally filtered and detected by a silicon single-photon avalanche photodiode, demonstrating the feasibility of an upconversion single-photon detector on-chip with a detection efficiency of 8.7% and a noise count rate of 300 cps. The realization of a low-noise QFC device paves the way for practical chip-scale QFC-based quantum systems in heterogeneous configurations.

quant-ph↗

arXiv2022

Constraining a modified gravity theory in strong gravitational lensing and black hole shadow observations

We study the strong gravitational lensing effect around rotating black holes in different gravity theories. By calculating the deflection angle of strong gravitational lensing, we evaluate the lensing observables including the image position, separation, magnification and the time delays between the relativistic images of different rotating black holes. We argue that the differences in image positions, separations between the rotating black hole in modified gravity (MOG) theory and the Kerr black hole in general relativity (GR) are more significant in SgrA* than those in M87*, however the differences in time delays between rotating black holes in MOG and GR are shorter in SgrA* than that in M87*. Our evaluations on lensing observables in the strong gravity regime can help to distinguish the MOG from GR. Furthermore, we investigate the shadow observables of different rotating black holes. Employing the EHT observations on the angular shadow radius for supermassive M87* and SgrA* black holes respectively, we estimate the ranges of MOG parameter and obtain its upper limit constraint $0.350\lesssimα_{\rm up}\lesssim 0.485$ and $0.162 \lesssim α_{\rm up} \lesssim 0.285$ correspondingly, relating to black hole spins. This is the first constraint on the MOG parameter for rotating supermassive black holes from EHT observations on the angular shadow radius. Our constraint on the MOG parameter is much tighter compared with the result obtained from the orbital precession of the S2 star.

gr-qc↗

arXiv2022

Yes, DLGM! A novel hierarchical model for hazard classification

Hazards can be exposed by HAZOP as text information, and studying their classification is of great significance to the development of industrial informatics, which is conducive to safety early warning, decision support, policy evaluation, etc. However, there is no research on this important field at present. In this paper, we propose a novel model termed DLGM via deep learning for hazard classification. Specifically, first, we leverage BERT to vectorize the hazard and treat it as a type of time series (HTS). Secondly, we build a grey model FSGM(1, 1) to model it, and get the grey guidance in the sense of the structural parameters. Finally, we design a hierarchical-feature fusion neural network (HFFNN) to investigate the HTS with grey guidance (HTSGG) from three themes, where, HFFNN is a hierarchical structure with four types of modules: two feature encoders, a gating mechanism, and a deepening mechanism. We take 18 industrial processes as application cases and launch a series of experiments. The experimental results prove that DLGM has promising aptitudes for hazard classification and that FSGM(1, 1) and HFFNN are effective. We hope our research can contribute added value and support to the daily practice in industrial safety.

cs.CL↗

arXiv2022

Testing synchrotron models and frequency resolution in BINGO 21 cm simulated maps using GNILC

To recover the 21 cm hydrogen line, it is essential to separate the cosmological signal from the much stronger foreground contributions at radio frequencies. The BINGO radio telescope is designed to measure the 21 cm line and detect BAOs using the intensity mapping technique. This work analyses the performance of the GNILC method, combined with a power spectrum debiasing procedure. The method was applied to a simulated BINGO mission, building upon previous work from the collaboration. It compares two different synchrotron emission models and different instrumental configurations, in addition to the combination with ancillary data to optimize both the foreground removal and recovery of the 21 cm signal across the full BINGO frequency band, as well as to determine an optimal number of frequency bands for the signal recovery. We have produced foreground emissions maps using the Planck Sky Model, the cosmological Hi emission maps are generated using the FLASK package and thermal noise maps are created according to the instrumental setup. We apply the GNILC method to the simulated sky maps to separate the Hi plus thermal noise contribution and, through a debiasing procedure, recover an estimate of the noiseless 21 cm power spectrum. We found a near optimal reconstruction of the Hi signal using a 80 bins configuration, which resulted in a power spectrum reconstruction average error over all frequencies of 3%. Furthermore, our tests showed that GNILC is robust against different synchrotron emission models. Finally, adding an extra channel with CBASS foregrounds information, we reduced the estimation error of the 21 cm signal. The optimisation of our previous work, producing a configuration with an optimal number of channels for binning the data, impacts greatly the decisions regarding BINGO hardware configuration before commissioning.

astro-ph.CO↗

arXiv2022

Flux Linkage Based Evaluation Method for Voltage Inertia and Voltage Recovery Capability Under Large Disturbances

High-voltage direct current (HVDC) transmission applications and the growth of the dynamic load in large-scale receiving-end grids lead to a higher risk of short-term voltage instability. An effective way to address this problem is to improve the system's dynamic voltage support capability by changing the operation status of dynamic var devices, and the dynamic var reserve (DVR) is commonly used. Due to the time delay in synchronous machine excitation systems, the dynamic var reserved at steady state cannot be exploited immediately under large disturbances. In addition, some reactive power is produced immediately through electromagnetic induction. The voltage support effect of the two capabilities is analyzed based on the flux linkage and an approximate simulation of the fault impact. Then two novel indexes for evaluating the voltage inertia and voltage recovery capability are proposed, which are related to the voltage nadir and voltage recovery speed. The indexes' physical meanings are revealed by comparison with the frequency response, and the potential applications in planning and optimal reactive power dispatch (ORPD) are introduced. Numerical simulations based on the IEEE 39-bus system verify that the indexes can quantify the voltage support capabilities, and the minimum voltage support requirements are obtained to maintain systems' security.

eess.SY↗

arXiv2022

Dynamical transitions in scalarization and descalarization through black hole accretion

We present the first fully nonlinear study on the accretion of scalar fields onto a seed black hole in anti-de Sitter spacetime in Einstein-Maxwell-scalar theory. Intrinsic critical phenomena in the dynamical transition between the bald and scalarized black holes are disclosed. In scalarizations, the transition is discontinuous and a metastable black hole acts as an attractor at the threshold. We construct a new physical mechanism to dynamically descalarize an isolated scalarized black hole. The first results on critical phenomena in descalarizations are revealed. The dynamical descalarizations can be either discontinuous or continuous at the threshold, distinguished by whether or not an attractor appears.

gr-qc↗

arXiv2022

The BINGO project VIII: On the recoverability of the BAO signal on HI intensity mapping simulations

A new and promising technique for observing the Universe and study the dark sector is the intensity mapping of the redshifted 21cm line of neutral hydrogen (HI). The BINGO radio telescope will use the 21cm line to map the Universe in the redshift range $0.127 \le z \le 0.449$, in a tomographic approach, with the main goal of probing BAO. This work presents the forecasts of measuring the transversal BAO signal during the BINGO Phase 1 operation. We use two clustering estimators, the two-point angular correlation function (ACF) and the angular power spectrum (APS), and a template-based method to model the ACF and APS estimated from simulations of the BINGO region and extract the BAO information. The tomographic approach allows the combination of redshift bins to improve the template fitting performance. We find that each clustering estimator shows different sensitivities to specific redshift ranges, although both of them perform better at higher redshifts. In general, the APS estimator provides slightly better estimates, with smaller uncertainties and larger probability of detection of the BAO signal, achieving $\gtrsim 90$\% at higher redshifts. We investigate the contribution from instrumental noise and residual foreground signals and find that the former has the greater impact, getting more significant as the redshift increases, in particular the APS estimator. Indeed, including noise in the analysis increases the uncertainty up to a factor of $\sim 2.2$ at higher redshifts. Foreground residuals, in contrast, do not significantly affect our final uncertainties. In summary, our results show that, even including semi-realistic systematic effects, BINGO has the potential to successfully measure the BAO scale in radio frequencies. (Abridged)

astro-ph.CO↗

arXiv2022

Towards the Desirable Decision Boundary by Moderate-Margin Adversarial Training

Adversarial training, as one of the most effective defense methods against adversarial attacks, tends to learn an inclusive decision boundary to increase the robustness of deep learning models. However, due to the large and unnecessary increase in the margin along adversarial directions, adversarial training causes heavy cross-over between natural examples and adversarial examples, which is not conducive to balancing the trade-off between robustness and natural accuracy. In this paper, we propose a novel adversarial training scheme to achieve a better trade-off between robustness and natural accuracy. It aims to learn a moderate-inclusive decision boundary, which means that the margins of natural examples under the decision boundary are moderate. We call this scheme Moderate-Margin Adversarial Training (MMAT), which generates finer-grained adversarial examples to mitigate the cross-over problem. We also take advantage of logits from a teacher model that has been well-trained to guide the learning of our model. Finally, MMAT achieves high natural accuracy and robustness under both black-box and white-box attacks. On SVHN, for example, state-of-the-art robustness and natural accuracy are achieved.

cs.CV↗

arXiv2022

CompoundE: Knowledge Graph Embedding with Translation, Rotation and Scaling Compound Operations

Translation, rotation, and scaling are three commonly used geometric manipulation operations in image processing. Besides, some of them are successfully used in developing effective knowledge graph embedding (KGE) models such as TransE and RotatE. Inspired by the synergy, we propose a new KGE model by leveraging all three operations in this work. Since translation, rotation, and scaling operations are cascaded to form a compound one, the new model is named CompoundE. By casting CompoundE in the framework of group theory, we show that quite a few scoring-function-based KGE models are special cases of CompoundE. CompoundE extends the simple distance-based relation to relation-dependent compound operations on head and/or tail entities. To demonstrate the effectiveness of CompoundE, we conduct experiments on three popular KG completion datasets. Experimental results show that CompoundE consistently achieves the state of-the-art performance.

cs.AI↗

arXiv2022

Calibrating the momentum width of a trapped Bose-Einstein condensate by one-dimensional optical lattice pulse sequences

We experimentally measured the ultra-narrow momentum width of an optical trapped Bose-Einstein condensate (BEC) in situ based on matter-wave interference, which validates our previous theoretical work [arXiv: 2205.02416]. By sweeping the interval of double stand-wave pulses, the BEC wave packet was splitted into different diffraction orders and then we counted the oscillation curve of the population of zero-momentum state to calibrate the momentum width. Compared with our simplified theory, we observed an accelerated evolution of interference fringes in time-domain. We evaluated this interference process minutely by numerically calculating the Gross-Pitaevskii equation and using Wigner function to intuitively demonstrate the influence of the external potential and nonlinear term. We confirmed that the reduction of interference fringe evolution period actually originates from the synergistic cooperation of the mean-field interaction of the BEC and spatial density modulation caused by the interference between different momentum states. Our approach could be generalized to other ultra-cold atomic gases with different momentum distributions, and in principle a single shot can obtain the result. This quantum thermometry is particularly suitable for momentum width calibration in practical deep cooling experiments, while for atomic samples at pK level the mean-field interaction can be safely ignored.

cond-mat.quant-gas↗

arXiv2022

Hessian-Free Second-Order Adversarial Examples for Adversarial Learning

Recent studies show deep neural networks (DNNs) are extremely vulnerable to the elaborately designed adversarial examples. Adversarial learning with those adversarial examples has been proved as one of the most effective methods to defend against such an attack. At present, most existing adversarial examples generation methods are based on first-order gradients, which can hardly further improve models' robustness, especially when facing second-order adversarial attacks. Compared with first-order gradients, second-order gradients provide a more accurate approximation of the loss landscape with respect to natural examples. Inspired by this, our work crafts second-order adversarial examples and uses them to train DNNs. Nevertheless, second-order optimization involves time-consuming calculation for Hessian-inverse. We propose an approximation method through transforming the problem into an optimization in the Krylov subspace, which remarkably reduce the computational complexity to speed up the training procedure. Extensive experiments conducted on the MINIST and CIFAR-10 datasets show that our adversarial learning with second-order adversarial examples outperforms other fisrt-order methods, which can improve the model robustness against a wide range of attacks.

cs.LG↗

arXiv2022

Continuous-stage symplectic adapted exponential methods for charged-particle dynamics with arbitrary electromagnetic fields

This paper is devoted to the numerical symplectic approximation of the charged-particle dynamics (CPD) with arbitrary electromagnetic fields. By utilizing continuous-stage methods and exponential integrators, a general class of symplectic methods is formulated for CPD under a homogeneous magnetic field. Based on the derived symplectic conditions, two practical symplectic methods up to order four are constructed where the error estimates show that the proposed second order scheme has a uniform accuracy in the position w.r.t. the strength of the magnetic field. Moreover, the symplectic methods are extended to CPD under non-homogeneous magnetic fields and three algorithms are formulated. Rigorous error estimates are investigated for the proposed methods and one method is proved to have a uniform accuracy in the position w.r.t. the strength of the magnetic field. Numerical experiments are provided for CPD under homogeneous and non-homogeneous magnetic fields, and the numerical results support the theoretical analysis and demonstrate the remarkable numerical behavior of our methods.

math.NA↗

arXiv2022

Rethinking Reinforcement Learning based Logic Synthesis

Recently, reinforcement learning has been used to address logic synthesis by formulating the operator sequence optimization problem as a Markov decision process. However, through extensive experiments, we find out that the learned policy makes decisions independent from the circuit features (i.e., states) and yields an operator sequence that is permutation invariant to some extent in terms of operators. Based on these findings, we develop a new RL-based method that can automatically recognize critical operators and generate common operator sequences generalizable to unseen circuits. Our algorithm is verified on both the EPFL benchmark, a private dataset and a circuit at industrial scale. Experimental results demonstrate that it achieves a good balance among delay, area and runtime, and is practical for industrial usage.

cs.LG↗