• A three-dimensional Dirac semimetal has bulk Dirac cones in all three momentum directions and Fermi arc-like surface states, and can be converted into a Weyl semimetal by breaking time-reversal symmetry. However, the highly conductive bulk state usually hides the electronic transport from the surface state in Dirac semimetal. Here, we demonstrate the supercurrent carried by bulk and surface states in Nb-Cd3As2 nanowire-Nb short and long junctions, respectively. For the 1 micrometer long junction, the Fabry-Perot interferences induced oscillations of the critical supercurrent are observed, suggesting the ballistic transport of the surface states carried supercurrent, where the bulk states are decoherent and the topologically protected surface states still keep coherent. Moreover, a superconducting dome is observed in the long junction, which is attributed to the enhanced dephasing from the interaction between surface and bulk states as tuning gate voltage to increase the carrier density. The superconductivity of topological semimetal nanowire is promising for braiding of Majorana fermions toward topological quantum computing.
  • Recurrent Neural Networks (RNNs) are becoming increasingly important for time series-related applications which require efficient and real-time implementations. The recent pruning based work ESE suffers from degradation of performance/energy efficiency due to the irregular network structure after pruning. We propose block-circulant matrices for weight matrix representation in RNNs, thereby achieving simultaneous model compression and acceleration. We aim to implement RNNs in FPGA with highest performance and energy efficiency, with certain accuracy requirement (negligible accuracy degradation). Experimental results on actual FPGA deployments shows that the proposed framework achieves a maximum energy efficiency improvement of 35.7$\times$ compared with ESE.
  • Recently, significant accuracy improvement has been achieved for acoustic recognition systems by increasing the model size of Long Short-Term Memory (LSTM) networks. Unfortunately, the ever-increasing size of LSTM model leads to inefficient designs on FPGAs due to the limited on-chip resources. The previous work proposes to use a pruning based compression technique to reduce the model size and thus speedups the inference on FPGAs. However, the random nature of the pruning technique transforms the dense matrices of the model to highly unstructured sparse ones, which leads to unbalanced computation and irregular memory accesses and thus hurts the overall performance and energy efficiency. In contrast, we propose to use a structured compression technique which could not only reduce the LSTM model size but also eliminate the irregularities of computation and memory accesses. This approach employs block-circulant instead of sparse matrices to compress weight matrices and reduces the storage requirement from $\mathcal{O}(k^2)$ to $\mathcal{O}(k)$. Fast Fourier Transform algorithm is utilized to further accelerate the inference by reducing the computational complexity from $\mathcal{O}(k^2)$ to $\mathcal{O}(k\text{log}k)$. The datapath and activation functions are quantized as 16-bit to improve the resource utilization. More importantly, we propose a comprehensive framework called C-LSTM to automatically optimize and implement a wide range of LSTM variants on FPGAs. According to the experimental results, C-LSTM achieves up to 18.8X and 33.5X gains for performance and energy efficiency compared with the state-of-the-art LSTM implementation under the same experimental setup, and the accuracy degradation is very small.
  • With the wide application of machine learning algorithms to the real world, class imbalance and concept drift have become crucial learning issues. Class imbalance happens when the data categories are not equally represented, i.e., at least one category is minority compared to other categories. It can cause learning bias towards the majority class and poor generalization. Concept drift is a change in the underlying distribution of the problem, and is a significant issue specially when learning from data streams. It requires learners to be adaptive to dynamic changes. Class imbalance and concept drift can significantly hinder predictive performance, and the problem becomes particularly challenging when they occur simultaneously. This challenge arises from the fact that one problem can affect the treatment of the other. For example, drift detection algorithms based on the traditional classification error may be sensitive to the imbalanced degree and become less effective; and class imbalance techniques need to be adaptive to changing imbalance rates, otherwise the class receiving the preferential treatment may not be the correct minority class at the current moment. Therefore, the mutual effect of class imbalance and concept drift should be considered during algorithm design. The aim of this workshop is to bring together researchers from the areas of class imbalance learning and concept drift in order to encourage discussions and new collaborations on solving the combined issue of class imbalance and concept drift. It provides a forum for international researchers and practitioners to share and discuss their original work on addressing new challenges and research issues in class imbalance learning, concept drift, and the combined issues of class imbalance and concept drift. The proceedings include 8 papers on these topics.
  • We report an observation of a topologically protected transport of surface carriers in a quasi-ballistic Cd3As2 nanowire.The nanowire is thin enough for the spin-textured surface carriers to form 1D subbands, demonstrating conductance oscillations with gate voltage even without magnetic field. The {\pi} phase-shift of Aharonov-Bohm oscillations can periodically appear or disappear by tuning gate voltage continuously. Such a {\pi} phase shift stemming from the Berry's phase demonstrates the topological nature of surface states.The topologically protected transport of the surface states is further revealed by four-terminal nonlocal measurements.
  • As the explosive growth of smart devices and the advent of many new applications, traffic volume has been growing exponentially. The traditional centralized network architecture cannot accommodate such user demands due to heavy burden on the backhaul links and long latency. Therefore, new architectures which bring network functions and contents to the network edge are proposed, i.e., mobile edge computing and caching. Mobile edge networks provide cloud computing and caching capabilities at the edge of cellular networks. In this survey, we make an exhaustive review on the state-of-the-art research efforts on mobile edge networks. We first give an overview of mobile edge networks including definition, architecture and advantages. Next, a comprehensive survey of issues on computing, caching and communication techniques at the network edge is presented respectively. The applications and use cases of mobile edge networks are discussed. Subsequently, the key enablers of mobile edge networks such as cloud technology, SDN/NFV and smart devices are discussed. Finally, open research challenges and future directions are presented as well.
  • Caching popular contents at the edge of cellular networks has been proposed to reduce the load, and hence the cost of backhaul links. It is significant to decide which files should be cached and where to cache them. In this paper, we propose a distributed caching scheme considering the tradeoff between the diversity and redundancy of base stations' cached contents. Whether it is better to cache the same or different contents in different base stations? To find out this, we formulate an optimal redundancy caching problem. Our goal is to minimize the total transmission cost of the network, including cost within the radio access network (RAN) and cost incurred by transmission to the core network via backhaul links. The optimal redundancy ratio under given system configuration is obtained with adapted particle swarm optimization (PSO) algorithm. We analyze the impact of important system parameters through Monte-Carlo simulation. Results show that the optimal redundancy ratio is mainly influenced by two parameters, which are the backhaul to RAN unit cost ratio and the steepness of file popularity distribution. The total cost can be reduced by up to 54% at given unit cost ratio of backhaul to RAN when the optimal redundancy ratio is selected. Under typical file request pattern, the reduction amount can be up to 57%.
  • The volume and types of traffic data in mobile cellular networks have been increasing continuously. Meanwhile, traffic data change dynamically in several dimensions such as time and space. Thus, traffic modeling is essential for theoretical analysis and energy efficient design of future ultra-dense cellular networks. In this paper, the authors try to build a tractable and accurate model to describe the traffic variation pattern for a single base station in real cellular networks. Firstly a sinusoid superposition model is proposed for describing the temporal traffic variation of multiple base stations based on real data in a current cellular network. It shows that the mean traffic volume of many base stations in an area changes periodically and has three main frequency components. Then, lognormal distribution is verified for spatial modeling of real traffic data. The spatial traffic distributions at both spare time and busy time are analyzed. Moreover, the parameters of the model are presented in three typical regions: park, campus and central business district. Finally, an approach for combined spatial-temporal traffic modeling of single base station is proposed based on the temporal and spatial traffic distribution of multiple base stations. All the three models are evaluated through comparison with real data in current cellular networks. The results show that these models can accurately describe the variation pattern of real traffic data in cellular networks.
  • As an emerging research topic, online class imbalance learning often combines the challenges of both class imbalance and concept drift. It deals with data streams having very skewed class distributions, where concept drift may occur. It has recently received increased research attention; however, very little work addresses the combined problem where both class imbalance and concept drift coexist. As the first systematic study of handling concept drift in class-imbalanced data streams, this paper first provides a comprehensive review of current research progress in this field, including current research focuses and open challenges. Then, an in-depth experimental study is performed, with the goal of understanding how to best overcome concept drift in online learning with class imbalance. Based on the analysis, a general guideline is proposed for the development of an effective algorithm.
  • Optimal control of bilinear systems has been a well-studied subject in the areas of mathematical and computational optimal control. However, effective methods for solving emerging optimal control problems involving an ensemble of deterministic or stochastic bilinear systems are underdeveloped. These burgeoning problems arise in diverse applications from quantum control and molecular imaging to neuroscience. In this work, we develop an iterative method to find optimal controls for an inhomogeneous bilinear ensemble system with free-endpoint conditions. The central idea is to represent the bilinear ensemble system at each iteration as a time-varying linear ensemble system, and then solve it in an iterative manner. We analyze convergence of the iterative procedure and discuss optimality of the convergent solutions. The method is directly applicable to solve the same class of optimal control problems involving a stochastic bilinear ensemble system driven by independent additive noise processes. We demonstrate the robustness and applicability of the developed iterative method through practical control designs in neuroscience and quantum control.
  • Multiple chiral doublet bands (M$\chi$D) in the $80$, 130 and $190$ mass regions are studied by the model of $\gamma$=90$^{\circ}$ triaxial rotor coupled with identical symmetric proton-neutron configurations. By selecting the suitable basis, the calculated wave functions are explicitly exhibited to be symmetric under the operator $\hat{A}$, which is defined as rotation by $90^{\circ}$ about 3-axis with the exchange of valance proton and neutron. We found that both $M1$ and $E2$ transitions are allowed between the levels with different values of $A$, while are forbidden between the levels with same values of $A$. Such a selection rule holds true for M$\chi$D in different mass regions.
  • We engineer the fast rotation of a quantum particle confined in an effectively one-dimensional, harmonic trap, for a predetermined rotation angle and time, avoiding final excitation. Different schemes are proposed with different speed limits that depend on the control capabilities. We also make use of trap rotations to create squeezed states without manipulating the trap frequencies.
  • The rotation of sunspots of 2 yr in two different solar cycles is studied with the data from the Helioseismic and Magnetic Imager on board the \it Solar Dynamics Observatory \rm and the Michelson Doppler Imager instrument on board the \it Solar and Heliospheric Observataory.\rm We choose the $\alpha$ sunspot groups and the relatively large and stable sunspots of complex active regions in our sample. In the year of 2003, the $\alpha$ sunspot groups and the preceding sunspots tend to rotate counterclockwise and have positive magnetic polarity in the northern hemisphere. In the southern hemisphere, the magnetic polarity and rotational tendency of the $\alpha$ sunspot groups and the preceding sunspots are opposite to the northern hemisphere. The average rotational speed of these sunspots in 2003 is about $0^{\circ}.65 \rm \ hr^{-1}$. From 2014 January to 2015 February, the $\alpha$ sunspot groups and the preceding sunspots tend to rotate clockwise and have negative magnetic polarity in the northern hemisphere. The patterns of rotation and magnetic polarity of the southern hemisphere are also opposite to those of the northern hemisphere. The average rotational speed of these sunspots in 2014/2015 is about $1^{\circ}.49 \rm \ hr^{-1}$. The rotation of the relatively large and stable preceding sunspots and that of the $\alpha$ sunspot groups located in the same hemisphere have opposite rotational direction in 2003 and 2014/2015.
  • Stochastic effect in cellular systems has been an important topic in systems biology. Stochastic modeling and simulation methods are important tools to study stochastic effect. Given the low efficiency of stochastic simulation algorithms, the hybrid method, which combines an ordinary differential equation (ODE) system with a stochastic chemically reacting system, shows its unique advantages in the modeling and simulation of biochemical systems. The efficiency of hybrid method is usually limited by reactions in the stochastic subsystem, which are modeled and simulated using Gillespie's framework and frequently interrupt the integration of the ODE subsystem. In this paper we develop an efficient implementation approach for the hybrid method coupled with traditional ODE solvers. We also compare the efficiency of hybrid methods with three widely used ODE solvers RADAU5, DASSL, and DLSODAR. Numerical experiments with three biochemical models are presented. A detailed discussion is presented for the performances of three ODE solvers.
  • Optimal control of bilinear systems has been a well-studied subject in the area of mathematical control. However, techniques for solving emerging optimal control problems involving an ensemble of structurally identical bilinear systems are underdeveloped. In this work, we develop an iterative method to effectively and systematically solve these challenging optimal ensemble control problems, in which the bilinear ensemble system is represented as a time-varying linear ensemble system at each iteration and the optimal ensemble control law is then obtained by the singular value expansion of the input-to-state operator that describes the dynamics of the linear ensemble system. We examine the convergence of the developed iterative procedure and pose optimality conditions for the convergent solution. We also provide examples of practical control designs in magnetic resonance to demonstrate the applicability and robustness of the developed iterative method.
  • Chromospheric rapid blueshifted excursions (RBEs) are suggested to be the disk counterparts of type II spicules at the limb and believed to contribute to the coronal heating process. Previous identification of RBEs was mainly based on feature detection using Dopplergrams. In this paper, we study RBEs on 2011 October 21 in a very quiet region at the disk center, which were observed with the high-cadence imaging spectroscopy of the Ca II 8542 A line from the Interferometric Bidimensional Spectrometer (IBIS). By using an automatic spectral analysis algorithm, a total of 98 RBEs are identified during a 11 minute period. Most of these RBEs have either a round or elongated shape, with an average area of 1.2 arcsec^2. The detailed temporal evolution of spectra from IBIS makes possible a quantitative determination of the velocity (~16 km/s) and acceleration (~400 m/s^2) of Ca II 8542 RBEs, and reveal an additional deceleration (~-160 m/s^2) phase that usually follows the initial acceleration. In addition, we also investigate the association of RBEs with the concomitant photospheric magnetic field evolution, using coordinated high-resolution and high-sensitivity magnetograms made by Hinode. Clear examples are found where RBEs appear to be associated with the preceding magnetic flux emergence and/or the subsequent flux cancellation. However, a further analysis with the aid of the Southwest Automatic Magnetic Identification Suite does not yield a significant statistical association between these RBEs and magnetic field evolution. We discuss the implications of our results in the context of understanding the driving mechanism of RBEs.
  • Large, complex, active regions may produce multiple flares within a certain period of one or two days. These flares could occur in the same location with similar morphologies, commonly referred to as homologous flares. In 2011 September, active region NOAA 11283 produced a pair of homologous flares on the 6th and 7th, respectively. Both of them were white-light (WL) flares, as captured by the Helioseismic and Magnetic Imager (HMI) onboard the Solar Dynamics Observatory in visible continuum at 617.3 nm which is believed to originate from the deep solar atmosphere.We investigate the WL emission of these X-class flares with HMIs seeing-free imaging spectroscopy. The durations of impulsive peaks in the continuum are about 4 minutes. We compare the WL with hard X-ray (HXR) observations for the September 6 flare and find a good correlation between the continuum and HXR both spatially and temporally. In absence of RHESSI data during the second flare on September 7, the derivative of the GOES soft X-ray is used and also found to be well correlated temporally with the continuum. We measure the contrast enhancements, characteristic sizes, and HXR fluxes of the twin flares, which are similar for both flares, indicating analogous triggering and heating processes. However, the September 7 flare was associated with conspicuous sunquake signals whereas no seismic wave was detected during the flare on September 6. Therefore, this comparison suggests that the particle bombardment may not play a dominant role in producing the sunquake events studied in this paper.
  • We present the observation of a major solar eruption that is associated with fast sunspot rotation. The event includes a sigmoidal filament eruption, a coronal mass ejection, and a GOES X2.1 flare from NOAA active region 11283. The filament and some overlying arcades were partially rooted in a sunspot. The sunspot rotated at $\sim$10$^\circ$ per hour rate during a period of 6 hours prior to the eruption. In this period, the filament was found to rise gradually along with the sunspot rotation. Based on the HMI observation, for an area along the polarity inversion line underneath the filament, we found gradual pre-eruption decreases of both the mean strength of the photospheric horizontal field ($B_h$) and the mean inclination angle between the vector magnetic field and the local radial (or vertical) direction. These observations are consistent with the pre-eruption gradual rising of the filament-associated magnetic structure. In addition, according to the Non-Linear Force-Free-Field reconstruction of the coronal magnetic field, a pre-eruption magnetic flux rope structure is found to be in alignment with the filament, and a considerable amount of magnetic energy was transported to the corona during the period of sunspot rotation. Our study provides evidences that in this event sunspot rotation plays an important role in twisting, energizing, and destabilizing the coronal filament-flux rope system, and led to the eruption. We also propose that the pre-event evolution of $B_h$ may be used to discern the driving mechanism of eruptions.
  • The Helioseismic and Magnetic Imager provides 45 s cadence intensity images and 720 s cadence vector magnetograms. These unprecedented high-cadence and high-resolution data give us a unique opportunity to study the change of photospheric flows and sunspot rotations associated with flares. By using the differential affine velocity estimator method and the Fourier local correlation tracking method separately, we calculate velocity and vorticity of photospheric flows in the flaring NOAA AR 11158, and investigate their temporal evolution around the X2.2 flare on 2011 February 15. It is found that the shear flow around the flaring magnetic polarity inversion line exhibits a sudden decrease, and both of the two main sunspots undergo a sudden change in rotational motion during the impulsive phase of the flare. These results are discussed in the context of the Lorentz-force change that was proposed by Hudson et al. (2008) and Fisher et al. (2012). This mechanism can explain the connections between the rapid and irreversible photospheric vector magnetic field change and the observed short-term motions associated with the flare. In particular, the torque provided by the horizontal Lorentz force change agrees with what is required for the measured angular acceleration.
  • Rapid, irreversible changes of magnetic topology and sunspot structure associated with flares have been systematically observed in recent years. The most striking features include the increase of horizontal field at the polarity inversion line (PIL) and the co-spatial penumbral darkening. A likely explanation of the above phenomenon is the back reaction to the coronal restructuring after eruptions: a coronal mass ejection carries the upward momentum while the downward momentum compresses the field lines near the PIL. Previous studies could only use low resolution (above 1") magnetograms and white-light images. Therefore, the changes are mostly observed for X-class flares. Taking advantage of the 0.1" spatial resolution and 15s temporal cadence of the New Solar Telescope at Big Bear Solar Observatory, we report in detail the rapid formation of sunspot penumbra at the PIL associated with the C7.4 flare on 2012 July 2. It is unambiguously shown that the solar granulation pattern evolves to alternating dark and bright fibril structure, the typical pattern of penumbra. Interestingly, the appearance of such a penumbra creates a new delta sunspot. The penumbral formation is also accompanied by the enhancement of horizontal field observed using vector magnetograms from the Helioseismic and Magnetic Imager. We explain our observations as due to the eruption of a flux rope following magnetic cancellation at the PIL. Subsequently the re-closed arcade fields are pushed down towards the surface to form the new penumbra. NLFFF extrapolation clearly shows both the flux rope close to the surface and the overlying fields.
  • The rapid and irreversible change of photospheric magnetic fields associated with flares has been confirmed by many recent studies. These studies showed that the photospheric magnetic fields respond to coronal field restructuring and turn to a more horizontal state near the magnetic polarity inversion line (PIL) after eruptions. Recent theoretical work has shown that the change in the Lorentz force associated with a magnetic eruption will lead to such a field configuration at the photosphere. The Helioseismic Magnetic Imager has been providing unprecedented full-disk vector magnetograms covering the rising phase of the solar cycle 24. In this study, we analyze 18 flares in four active regions, with GOES X-ray class ranging from C4.7 to X5.4. We find that there are permanent and rapid changes of magnetic field around the flaring PIL, the most notable of which is the increase of the transverse magnetic field. The changes of fields integrated over the area and the derived change of Lorentz force both show a strong correlation with flare magnitude. It is the first time that such magnetic field changes have been observed even for C-class flares. Furthermore, for seven events with associated CMEs, we use an estimate of the impulse provided by the Lorentz force, plus the observed CME velocity, to estimate the CME mass. We find that if the time scale of the back reaction is short, i.e., in the order of 10 s, the derived values of CME mass (10^{15} g) generally agree with those reported in literature.
  • It is well known that the long-term evolution of the photospheric magnetic field plays an important role in building up free energy to power solar eruptions. Observations, despite being controversial, have also revealed a rapid and permanent variation of the photospheric magnetic field in response to the coronal magnetic field restructuring during the eruption. The Helioseismic and Magnetic Imager instrument (HMI) on board the newly launched Solar Dynamics Observatory (SDO) produces seeing-free full-disk vector magnetograms at consistently high resolution and high cadence, which finally makes possible an unambiguous and comprehensive study of this important back-reaction process. In this study, we present a near disk-center, GOES -class X2.2 flare, which occurred in NOAA AR 11158 on 2011 February 15. Using the magnetic field measurements made by HMI, we obtained the first solid evidence of a rapid (in about 30 minutes) and irreversible enhancement in the horizontal magnetic field at the flaring magnetic polarity inversion line (PIL) by a magnitude of ~30%. It is also shown that the photospheric field becomes more sheared and more inclined. This field evolution is unequivocally associated with the flare occurrence in this sigmoidal active region, with the enhancement area located in between the two chromospheric flare ribbons and the initial conjugate hard X-ray footpoints. These results strongly corroborate our previous conjecture that the photospheric magnetic field near the PIL must become more horizontal after eruptions, which could be related to the newly formed low-lying fields resulted from the tether-cutting reconnection.
  • The rapid, irreversible change of the photospheric magnetic field has been recognized as an important element of the solar flare process. This Letter reports such a rapid change of magnetic fields during the 2011 February 13 M6.6 flare in NOAA AR 11158 that we found from the vector magnetograms of the Helioseismic and Magnetic Imager with 12-min cadence. High-resolution magnetograms of Hinode that are available at ~-5.5, -1.5, 1.5, and 4 hrs relative to the flare maximum are used to reconstruct three-dimensional coronal magnetic field under the nonlinear force-free field (NLFFF) assumption. UV and hard X-ray images are also used to illuminate the magnetic field evolution and energy release. The rapid change is mainly detected by HMI in a compact region lying in the center of the magnetic sigmoid, where the mean horizontal field strength exhibited a significant increase by 28%. The region lies between the initial strong UV and hard X-ray sources in the chromosphere, which are cospatial with the central feet of the sigmoid according to the NLFFF model. The NLFFF model further shows that strong coronal currents are concentrated immediately above the region, and that more intriguingly, the coronal current system underwent an apparent downward collapse after the sigmoid eruption. These results are discussed in favor of both the tether-cutting reconnection producing the flare and the ensuing implosion of the coronal field resulting from the energy release.
  • The commonly observed jets provide critical information on the small-scale energy release in the solar atmosphere. We report a near disk-center jet on 2010 July 20, observed by the Solar Dynamics Observatory. In this event, the standard interchange magnetic reconnection between an emerging flux spanning 9 x 10^3 km and ambient open fields is followed by a blowout-like eruption. In the "standard" stage, as the emerging negative element approached the nearby positive network fields, a jet with a dome-like base in EUV grew for 30 minutes before the jet spire began to migrate laterally with enhanced flux emergence. In the "blowout" stage, the above converging fields collided and the subsequent cancellation produced a UV microflare lasting seven minutes, in which the dome of the jet seemed to be blown out as (1) the spire swung faster and exhibited an unwinding motion before vanishing, (2) a rising loop and a blob erupted leaving behind cusped structures, with the blob spiraling outward in acceleration after the flare maximum, and (3) ejecting material with a curtain-like structure at chromospheric to transition-region temperatures also underwent a transverse motion. It is thus suggested that the flare reconnection rapidly removes the outer fields of the emerging flux to allow its twisted core field to erupt, a scenario favoring the jet-scale magnetic breakout model as recently advocated by Moore et al. in 2010.
  • Photospheric magnetic field not only plays important roles in building up free energy and triggering solar eruptions, but also has been observed to change rapidly and permanently responding to the coronal magnetic field restructuring due to coronal transients. The Helioseismic and Magnetic Imager instrument (HMI) on board the newly launched Solar Dynamics Observatory (SDO) produces seeing-free full-disk vector magnetograms at consistently high resolution and high cadence, which finally makes possible an unambiguous and comprehensive study of this important back-reaction process. In this study, we present a near disk-center, GOES-class X2.2 flare occurred at NOAA AR 11158 on 2011 February 15 using the magnetic field measurements made by HMI. We obtained the first solid evidence of an enhancement in the transverse magnetic field at the flaring magnetic polarity inversion line (PIL) by a magnitude of 70%. This rapid and irreversible field evolution is unequivocally associated with the flare occurrence, with the enhancement area located in between the two chromospheric flare ribbons. Similar findings have been made for another two major flare events observed by SDO. These results strongly corroborate our previous suggestion that the photospheric magnetic field near the PIL must become more horizontal after eruptions. In-depth studies will follow to further link the photospheric magnetic field changes with the dynamics of coronal mass ejections, when full Stokes inversion is carried out to generate accurate magnetic field vectors.