arXiv subjects

Chen Shi

Publications and source records attributed to Chen Shi.

At least 73 records · Page 4Linked to original sources

arXiv2022

Conservation of Total Wave Action in the Expanding Solar Wind

The conservation of wave action in moving plasmas has been well-known for over half a century. However, wave action is not conserved when multiple wave modes propagate and coexist close to degeneration condition (Sound speed equals Alfvén speed, i.e. plasma $β\sim 1$). Here we show that the violation of conservation is due to wave mode conversion, and that the total wave action summed over interacting modes is still conserved. Though the result is general, we focus on MHD waves and identify three distinctive mode conversion mechanisms, i.e. degeneracy, linear mode conversion, and resonance, and provide an intuitive physical picture for the mode conversion processes. We use 1D MHD simulations with the Expanding Box Model to simulate the nonlinear evolution of monochromatic MHD waves in the expanding solar wind. Simulation results validate the theory; total wave action therefore remains an interesting diagnostic for studies of waves and turbulence in the solar wind.

astro-ph.SR↗

arXiv2022

Magnetic field intermittency in the solar wind: PSP and SolO observations ranging from the Alfven region out to 1 AU

$PSP$ and $SolO$ data are utilized to investigate magnetic field intermittency in the solar wind (SW). Small-scale intermittency $(20-100d_{i})$ is observed to radially strengthen when methods relying on higher-order moments are considered ($SF_q$, $SDK$), but no clear trend is observed at larger scales. However, lower-order moment-based methods (e.g., PVI) are deemed more appropriate for examining the evolution of the bulk of Coherent Structures (CSs), $PVI \ge 3$. Using PVI, we observe a scale-dependent evolution in the fraction of the dataset occupied by CSs, $f_{PVI \ge 3}$. Specifically, regardless of the SW speed, a subtle increase is found in $f_{PVI\ge3}$ for $\ell =20 d_i$, in contrast to a more pronounced radial increase in CSs observed at larger scales. Intermittency is investigated in relation to plasma parameters. Though, slower SW speed intervals exhibit higher $f_{PVI \geq 6}$ and higher kurtosis maxima, no statistical differences are observed for $f_{PVI \geq 3}$. Highly Alfvénic intervals, display lower levels of intermittency. The anisotropy with respect to the angle between the magnetic field and SW flow, $Θ_{VB}$ is investigated. Intermittency is weaker at $Θ_{VB} \approx 0^{\circ}$ and is strengthened at larger angles. Considering the evolution at a constant alignment angle, a weakening of intermittency is observed with increasing advection time of the SW. Our results indicate that the strengthening of intermittency in the inner heliosphere is driven by the increase in comparatively highly intermittent perpendicular intervals sampled by the probes with increasing distance, an effect related directly to the evolution of the Parker spiral.

astro-ph.SR↗

arXiv2022

Influence of the heliospheric current sheet on the evolution of solar wind turbulence

The effects of the heliospheric current sheet (HCS) on the evolution of Alfvénic turbulence in the solar wind are studied using MHD simulations incorporating the expanding-box-model (EBM). The simulations show that near the HCS, the Alfvénicity of the turbulence decreases as manifested by lower normalized cross helicity and larger excess of magnetic energy. The numerical results are supported by a superposed-epoch analysis using OMNI data, which shows that the normalized cross helicity decreases inside the plasma sheet surrounding HCS, and the excess of magnetic energy is significantly enhanced at the center of HCS. Our simulation results indicate that the decrease of Alfvénicity around the HCS is due to the weakening of radial magnetic field and the effects of the transverse gradient in the background magnetic field. The magnetic energy excess in the turbulence may be a result of the loss of Alfvénic correlation between velocity and magnetic field and the faster decay of transverse kinetic energy with respect to magnetic energy in a spherically expanding solar wind.

astro-ph.SR↗

arXiv2022

The role of Alfvén wave dynamics on the large scale properties of the solar wind: comparing an MHD simulation with PSP E1 data

During Parker Solar Probe's first orbit, the solar wind plasma has been observed in situ closer than ever before, the perihelion on November 6th 2018 revealing a flow that is constantly permeated by large amplitude Alfvénic fluctuations. These include radial magnetic field reversals, or switchbacks, that seem to be a persistent feature of the young solar wind. The measurements also reveal a very strong, unexpected, azimuthal velocity component. In this work, we numerically model the solar corona during this first encounter, solving the MHD equations and accounting for Alfvén wave transport and dissipation. We find that the large scale plasma parameters are well reproduced, allowing the computation of the solar wind sources at Probe with confidence. We try to understand the dynamical nature of the solar wind to explain both the amplitude of the observed radial magnetic field and of the azimuthal velocities.

astro-ph.SR↗

arXiv2022

Statistical analysis of intermittency and its association with proton heating in the near Sun environment

We use data from the first six encounters of Parker Solar Probe and employ the Partial Variance of Increments ($PVI$) method to study the statistical properties of coherent structures in the inner heliosphere with the aim of exploring physical connections between magnetic field intermittency and observable consequences such as plasma heating and turbulence dissipation. Our results support proton heating localized in the vicinity of, and strongly correlated with, magnetic structures characterized by $PVI \geq 1$. We show that on average, such events constitute $\approx 19\%$ of the dataset, though variations may occur depending on the plasma parameters. We show that the waiting time distribution ($WT$) of identified events is consistent across all six encounters following a power-law scaling at lower $WTs$. This result indicates that coherent structures are not evenly distributed in the solar wind but rather tend to be tightly correlated and form clusters. We observe that the strongest magnetic discontinuities, $PVI \geq 6$, usually associated with reconnection exhausts, are sites where magnetic energy is locally dissipated in proton heating and are associated with the most abrupt changes in proton temperature. However, due to the scarcity of such events, their relative contribution to energy dissipation is minor. Taking clustering effects into consideration, we show that smaller scale, more frequent structures with PVI between, $1\lesssim PVI \lesssim 6$, play the major role in magnetic energy dissipation. The number density of such events is strongly associated with the global solar wind temperature, with denser intervals being associated with higher $T_{p}$.

astro-ph.SR↗

arXiv2021

NER-BERT: A Pre-trained Model for Low-Resource Entity Tagging

Named entity recognition (NER) models generally perform poorly when large training datasets are unavailable for low-resource domains. Recently, pre-training a large-scale language model has become a promising direction for coping with the data scarcity issue. However, the underlying discrepancies between the language modeling and NER task could limit the models' performance, and pre-training for the NER task has rarely been studied since the collected NER datasets are generally small or large but with low quality. In this paper, we construct a massive NER corpus with a relatively high quality, and we pre-train a NER-BERT model based on the created dataset. Experimental results show that our pre-trained model can significantly outperform BERT as well as other strong baselines in low-resource scenarios across nine diverse domains. Moreover, a visualization of entity representations further indicates the effectiveness of NER-BERT for categorizing a variety of entities.

cs.CL↗

arXiv2021

Stability of the magnetotail current sheet with normal magnetic field and field-aligned plasma flows

One of the most important problems of magnetotail dynamics is the substorm onset and the related instability of the magneotail current sheet. Since the simplest 2D current sheet configuration with monotonic $B_z$ was proven to be stable to the tearing mode, the focus of the instability investigation moved to more specific configurations, e.g. kinetic current sheets with strong transient ion currents and current sheets with non-monotonic $B_z$ (local $B_z$ minima or/and peaks). Stability of the latter current sheet configuration has been studied both within kinetic and fluid approaches, whereas the investigation of the transient ion effects were limited to kinetic models only. This paper aims to provide detailed analysis of stability of a multi-fluid current sheet configuration that mimics current sheets with transient ions. Using the system with two field-aligned ion flows that mimic the effect of pressure non-gyrotropy, we construct 1D current sheet with a finite $B_z$. This model describes well recent findings of very thin intense magnetotail current sheets. The stability analysis of this two-ion model confirms the stabilizing effect of finite $B_z$ and shows that the most stable current sheet is the one with exactly counter-streaming ion flows and zero net flow. Such field-aligned flows may substitute the contribution of the pressure tensor nongyrotropy to the stress balance, but cannot overtake the stabilizing effect of $B_z$. Obtained results are discussed in the context of magnetotail dynamical models and spacecraft observations.

physics.space-ph↗

arXiv2021

Evolution of switchbacks in the inner Heliosphere

We analyze magnetic field data from the first six encounters of PSP, three Helios fast streams and two Ulysses south polar passes covering heliocentric distances $0.1\lesssim R\lesssim 3$ au. We use this data set to statistically determine the evolution of switchbacks of different periods and amplitudes with distance from the Sun. We compare the radial evolution of magnetic field variances with that of the mean square amplitudes of switchbacks, and quantify the radial evolution of the cumulative counts of switchbacks per km. We find that the amplitudes of switchbacks decrease faster than the overall turbulent fluctuations, in a way consistent with the radial decrease of the mean magnetic field. This could be the result of a saturation of amplitudes and may be a signature of decay processes of large amplitude Alfvénic fluctuations in the solar wind. We find that the evolution of switchback occurrence in the solar wind is scale-dependent: the fraction of longer duration switchbacks increases with radial distance whereas it decreases for shorter switchbacks. This implies that switchback dynamics is a complex process involving both decay and in-situ generation in the inner heliosphere. We confirm that switchbacks can be generated by the expansion although other type of switchbacks generated closer to the sun cannot be ruled out.

astro-ph.SR↗

arXiv2021

BiToD: A Bilingual Multi-Domain Dataset For Task-Oriented Dialogue Modeling

Task-oriented dialogue (ToD) benchmarks provide an important avenue to measure progress and develop better conversational agents. However, existing datasets for end-to-end ToD modeling are limited to a single language, hindering the development of robust end-to-end ToD systems for multilingual countries and regions. Here we introduce BiToD, the first bilingual multi-domain dataset for end-to-end task-oriented dialogue modeling. BiToD contains over 7k multi-domain dialogues (144k utterances) with a large and realistic bilingual knowledge base. It serves as an effective benchmark for evaluating bilingual ToD systems and cross-lingual transfer learning approaches. We provide state-of-the-art baselines under three evaluation settings (monolingual, bilingual, and cross-lingual). The analysis of our baselines in different settings highlights 1) the effectiveness of training a bilingual ToD system compared to two independent monolingual ToD systems, and 2) the potential of leveraging a bilingual knowledge base and cross-lingual transfer learning to improve the system performance under low resource condition.

cs.CL↗

arXiv2021

Fast and Accurate Scene Parsing via Bi-direction Alignment Networks

In this paper, we propose an effective method for fast and accurate scene parsing called Bidirectional Alignment Network (BiAlignNet). Previously, one representative work BiSeNet~\cite{bisenet} uses two different paths (Context Path and Spatial Path) to achieve balanced learning of semantics and details, respectively. However, the relationship between the two paths is not well explored. We argue that both paths can benefit each other in a complementary way. Motivated by this, we propose a novel network by aligning two-path information into each other through a learned flow field. To avoid the noise and semantic gaps, we introduce a Gated Flow Alignment Module to align both features in a bidirectional way. Moreover, to make the Spatial Path learn more detailed information, we present an edge-guided hard pixel mining loss to supervise the aligned learning process. Our method achieves 80.1\% and 78.5\% mIoU in validation and test set of Cityscapes while running at 30 FPS with full resolution inputs. Code and models will be available at \url{https://github.com/jojacola/BiAlignNet}.

cs.CV↗

arXiv2021

Dynamic Dual Sampling Module for Fine-Grained Semantic Segmentation

Representation of semantic context and local details is the essential issue for building modern semantic segmentation models. However, the interrelationship between semantic context and local details is not well explored in previous works. In this paper, we propose a Dynamic Dual Sampling Module (DDSM) to conduct dynamic affinity modeling and propagate semantic context to local details, which yields a more discriminative representation. Specifically, a dynamic sampling strategy is used to sparsely sample representative pixels and channels in the higher layer, forming adaptive compact support for each pixel and channel in the lower layer. The sampled features with high semantics are aggregated according to the affinities and then propagated to detailed lower-layer features, leading to a fine-grained segmentation result with well-preserved boundaries. Experiment results on both Cityscapes and Camvid datasets validate the effectiveness and efficiency of the proposed approach. Code and models will be available at \url{x3https://github.com/Fantasticarl/DDSM}.

cs.CV↗

arXiv2021

Alfvénic versus non-Alfvénic turbulence in the inner heliosphere as observed by Parker Solar Probe

We make use of the Parker Solar Probe (PSP) data to explore the nature of solar wind turbulence focusing on the Alfvénic character and power spectra of the fluctuations and their dependence on distance and context (i.e. large scale solar wind properties), aiming to understand the role that different effects such as source properties, solar wind expansion, stream interaction might play in determining the turbulent state. We carry out a statistical survey of the data from the first five orbits of PSP with a focus on how the fluctuation properties at the large, MHD scales, vary with different solar wind streams and distance from the Sun. A more in-depth analysis from several selected periods is also presented. Our results show that as fluctuations are transported outward by the solar wind, the magnetic field spectrum steepens while the shape of the velocity spectrum remains unchanged. The steepening process is controlled by the "age" of the turbulence, determined by the wind speed together with the radial distance. Statistically, faster solar wind has higher "Alfvénicity", with more dominant outward propagating wave component and more balanced magnetic/kinetic energies. The outward wave dominance gradually weakens with radial distance, while the excess of magnetic energy is found to be stronger as we move closer toward the Sun. We show that the turbulence properties can vary significantly stream to stream even if these streams are of similar speed, indicating very different origins of these streams. Especially, the slow wind that originates near the polar coronal holes has much lower Alfvénicity compared with the slow wind that originates from the active regions/pseudostreamers. We show that structures such as heliospheric current sheets and velocity shears can play an important role in modifying the properties of the turbulence.

astro-ph.SR↗

arXiv2020

Oblique tearing mode instability: guide field and Hall effect

The tearing mode instability is one important mechanism that may explain the triggering of fast magnetic reconnection in astrophysical plasmas such as the solar corona and the Earth's magnetosphere. In this paper, the linear stability analysis of the tearing mode is carried out for a current sheet in the presence of a guide field, including the Hall effect. We show that the presence of a strong guide field does not modify the most unstable mode in the two-dimensional wave vector space orthogonal to the current gradient direction, which remains the fastest growing parallel mode. With the Hall effect, the inclusion of a guide field turns the non-dispersive propagation along the guide field direction to a dispersive one. The oblique modes have a wave-like structure along the normal direction of the current sheet and a strong guide field suppresses this structure while making the eigen-functions asymmetric.

physics.plasm-ph↗

arXiv2020

Tearing instability and periodic density perturbations in the slow solar wind

In contrast with the fast solar wind, that originates in coronal holes, the source of the slow solar wind is still debated. Often intermittent and enriched with low FIP elements -- akin to what is observed in closed coronal loops -- the slow wind could form in bursty events nearby helmet streamers. Slow winds also exhibit density perturbations which have been shown to be periodic and could be associated with flux ropes ejected from the tip of helmet streamers, as shown recently by the WISPR white light imager onboard Parker Solar Probe (PSP). In this work, we propose that the main mechanism controlling the release of flux ropes is a flow-modified tearing mode at the heliospheric current sheet (HCS). We use MHD simulations of the solar wind and corona to reproduce realistic configurations and outflows surrounding the HCS. We find that this process is able to explain long ($\sim 10-20$h) and short ($\sim 1-2$h) timescales of density structures observed in the slow solar wind. This study also sheds new light on the structure, topology and composition of the slow solar wind, and could be, in the near future, compared with white light and in situ PSP observations.

astro-ph.SR↗

arXiv2019

Propagation of Alfvén waves in the expanding solar wind with the fast-slow stream interaction

We carry out two-dimensional magnetohydrodynamic (MHD) simulations of an ensemble of Alfvénic fluctuations propagating in a structured, expanding solar wind including the presence of fast and slow solar wind streams. Using an appropriate expanding box model, the simulations incorporate the effects of fast-slow stream shear and compression and rarefaction self-consistently. We investigate the radial and longitudinal evolution of the cross-helicity, the total and residual energies and the power spectra of outward and inward Alfvénic fluctuations. The stream interaction is found to strongly affect the radial evolution of Alfvénic turbulence. The total energy in the Alfvén waves is depleted within the velocity shear regions, accompanied by the decrease of the normalized cross-helicity. The presence of stream-compression facilitates this process. Residual energy fluctuates around zero due to the correlation and de-correlation between the inward/outward waves but no net growth or decrease of the residual energy is observed. The radial power spectra of the inward/outward Alfvén waves show significant longitudinal variations. Kolmogorov-like spectra are developed only inside the fast and slow streams and when both the compression and shear are present. On the other hand, the spectra along the longitudinal direction show clear Kolmogorov-like inertial ranges in all cases.

astro-ph.SR↗

arXiv2019

Magnetic field kinks and folds in the solar wind

Parker Solar Probe (PSP) observations during its first encounter at 35.7 $R_\odot$ have shown the presence of magnetic field lines which are strongly perturbed to the point that they produce local inversions of the radial magnetic field, known as switchbacks. Their counterparts in the solar wind velocity field are local enhancements in the radial speed, or jets, displaying (in all components) the velocity-magnetic field correlation typical of large amplitude Alfvén waves propagating away from the Sun. Switchbacks and radial jets have previously been observed over a wide range of heliocentric distances by Helios, WIND and Ulysses, although they were prevalent in significantly faster streams than seen at PSP. Here we study via numerical MHD simulations the evolution of such large amplitude Alfvénic fluctuations by including, in agreement with observations, both a radial magnetic field inversion and an initially constant total magnetic pressure. Despite the extremely large excursion of magnetic and velocity fields, switchbacks are seen to persist for up to hundreds of Alfvén crossing times before eventually decaying due to the parametric decay instability. Our results suggest that such switchback/jet configurations might indeed originate in the lower corona and survive out to PSP distances, provided the background solar wind is sufficiently calm, in the sense of not being pervaded by strong density fluctuations or other gradients, such as stream or magnetic field shears, that might destabilize or destroy them over shorter timescales.

physics.space-ph↗

arXiv2019

Deep RNN Framework for Visual Sequential Applications

Extracting temporal and representation features efficiently plays a pivotal role in understanding visual sequence information. To deal with this, we propose a new recurrent neural framework that can be stacked deep effectively. There are mainly two novel designs in our deep RNN framework: one is a new RNN module called Context Bridge Module (CBM) which splits the information flowing along the sequence (temporal direction) and along depth (spatial representation direction), making it easier to train when building deep by balancing these two directions; the other is the Overlap Coherence Training Scheme that reduces the training complexity for long visual sequential tasks on account of the limitation of computing resources. We provide empirical evidence to show that our deep RNN framework is easy to optimize and can gain accuracy from the increased depth on several visual sequence problems. On these tasks, we evaluate our deep RNN framework with 15 layers, 7* than conventional RNN networks, but it is still easy to train. Our deep framework achieves more than 11% relative improvements over shallow RNN models on Kinetics, UCF-101, and HMDB-51 for video classification. For auxiliary annotation, after replacing the shallow RNN part of Polygon-RNN with our 15-layer deep CBM, the performance improves by 14.7%. For video future prediction, our deep RNN improves the state-of-the-art shallow model's performance by 2.4% on PSNR and SSIM. The code and trained models are published accompanied by this paper: https://github.com/BoPang1996/Deep-RNN-Framework.

cs.CV↗

arXiv2019

Fast Recursive Reconnection and the Hall effect: Hall-MHD Simulations

Magnetohydrodynamic (MHD) theory and simulations have shown that reconnection is triggered via a fast "ideal" tearing instability in current sheets whose inverse aspect ratio decreases to $a/L\sim S^{-1/3}$, with $S$ is the Lundquist number defined by the half-length $L$ of the current sheet (of thickness $2a$). Ideal tearing, in 2D sheets, triggers a hierarchical collapse via stretching of X-points and recursive instability. At each step, the local Lundquist number decreases, until the subsequent sheet thickness either approaches kinetic scales or the Lundquist number becomes sufficiently small. Here we carry out a series of Hall-MHD simulations to show how the Hall effect modifies recursive reconnection once the ion inertial scale is approached. We show that as the ion inertial length becomes of the order of the inner, singular layer thickness at some step of the recursive collapse, reconnection transits from the plasmoid-dominant regime to an intermediate plasmoid+Hall regime and then to the Hall-dominant regime. The structure around the X-point, the reconnection rate, the dissipation property and the power spectra are also modified significantly by the Hall effect.

astro-ph.SR↗