Search

Shiyuan Huang, Zesen Wang, Xudong Sun, Kuanxin Li The First Affiliated Hospital of Bengbu Medical University, Bengbu Medical University, Bengbu, Anhui Province, 233044, People’s Republic of ChinaCorrespondence: Kuanxin Li, The First Affiliated Hospital of Bengbu Medical University, Bengbu Medical University, Bengbu, Anhui Province, 233044, People’s Republic of China, Email kuanxinli@126.comBackground: Bone tissue engineering (BTE) is a promising alternative to autologous bone grafting for the clinical treatment of bone defects, and inorganic/organic composite hydrogels as BTE scaffolds are a hot spot in current research. The construction of nano-hydroxyapatite/gelatin methacrylate/oxidized sodium alginate (nHAP/GelMA/OSA), abbreviated as HGO, composite hydrogels loaded with bone morphogenetic protein 7 (BMP7) will provide a suitable 3D microenvironment to promote cell aggregation, proliferation, and differentiation, thus facilitating bone repair and regeneration.Methods: Dually-crosslinked hydrogels were fabricated by combining GelMA and OSA, while HGO hydrogels were formulated by incorporating varying amounts of nHAP. The hydrogels were physically and chemically characterized followed by the assessment of their biocompatibility. BMP7-HGO (BHGO) hydrogels were fabricated by incorporating suitable concentrations of BMP7 into HGO hydrogels. The osteogenic potential of BHGO hydrogels was then validated through in vitro experiments and using rat femoral defect models.Results: The addition of nHAP significantly improved the physical properties of the hydrogel, and the composite hydrogel with 10% nHAP demonstrated the best overall performance among all groups. The selected concentration of HGO hydrogel served as a carrier for BMP7 loading and was evaluated for its osteogenic potential both in vivo and in vitro. The BHGO hydrogel demonstrated superior in vitro osteogenic induction and in vivo potential for repairing bone tissue compared to the outcomes observed in the

The utilization of cation substitution presents a prospective approach to manipulate the structural characteristics and enhance the electrochemical functionality of spinel cobaltous sulfide (Co3S4). However, the underlying mechanism behind the impact of distinct cation substitutions on this phenomenon remains inadequately elucidated. In this study, we perform a thorough assessment to elucidate the influence of replacing cations on the pseudocapacitive properties of porous nanowires made of spinel bimetallic sulfide (MexCo3-xS4; Me=Mn, Ni, Cu, and Co). One of the top competitors, NiCo2S4, demonstrates a significant specific capacitance of 1032.7 F g−1 at a current density of 2 A g−1. Furthermore, it demonstrates an impressive capacitance retention rate of 92.1% after undergoing 8000 cycles. Moreover, the use of NiCo2S4 and AC as the anode and cathode in the hybrid supercapacitor (HSC) lead to a significant energy density of 49.3 Wh kg−1 at 1600 W kg−1, validating the effectiveness of the prepared porous nanowire-like NiCo2S4 as an appropriate substance for energy storage systems. Density functional theory (DFT) confirms that the substitution of cation can stimulate the electrochemical activity of Co, facilitate stronger inter-element interactions, and synergistically enhance the conductivity of cobalt-based bimetallic sulfides. Graphical abstract: The regulatory mechanism of cation substitution on the pseudocapacitance performance of MexCo3-xS4 is elucidated through the integration of DFT calculations and electrochemical analysis. (Figure presented.) © The Author(s), under exclusive licence to Springer Nature Switzerland AG 2024.

The pursuit of achieving wideband electromagnetic wave absorption to meet the demands in multiple fields has attracted attention in the field of electromagnetic wave absorption. Considering that the dipole moments of different scales can correspond to different polarization bands, the coupling of multi-scale loss mechanisms may help to realize the electromagnetic wave absorption in the wide band. Therefore, we fabricated N-doped Co/C composites derived from ZIF-67 through a facile wet chemistry approach and pyrolysis process. Due to the precise fabrication method and the design of the Stranski-Krastanow growth mode, the prepared N-doped Co/C composites contain abundant functional groups, magnetic particles, and cobalt quantum dots. By combining atomic-scale doping loss and mesoscopic-scale hollow structures, the Co/C composite materials achieve the coupling of multiple-scale loss mechanisms and demonstrate outstanding electromagnetic wave absorption performance: the minimum reflection loss (RLmin) value reaches −57.0 dB at 17.9 GHz when the thickness is 1.73 mm, and the effective absorption bandwidth (EABD) reaches the maximum value of 5.7 GHz when the thickness is 2 mm. To sum up, this work provides a new strategy for realizing the broadband absorption of further electromagnetic wave absorption materials. © The Author(s), under exclusive licence to Springer Nature Switzerland AG 2024.

It is an urgent problem to realize reliable microwave absorption materials (MAMs) with low density. To address this issue, a series of controlled experiments w ere carried out, which indicated that the tubular structure enables excellent microwave absorption properties with a lower powder filling rate. This performance is attributable to the combined dielectric and magnetic loss mechanisms provided by Co/C and the interface polarization facilitated by multiple heterogeneous interfaces. Particularly, Co@C nanotubes, benefiting from the enhanced heterointerface polarization due to their abundant specific surface area and the reduced electron migration barrier induced by their 1D stacked structure, effectively achieved a dual enhancement of dielectric loss and polarization loss at lower powder filling ratios. Furthermore, the magnetic coupling effect of magnetic nanoparticle arrays in tubular structures is demonstrated by micromagnetic simulation, which have been few reported elsewhere. These propertied enable Co@C nanotubes to achieve minimum reflection loss and maximum effective absorption broadband values of 61.0 dB and 5.5 GHz, respectively, with a powder filling ratio of 20 wt% and a thickness of 1.94 mm. This study reveals the significance of designing 1D structures in reducing powder filling ratio and matching thickness, providing valuable insights for developing MAMs with different microstructures. Hollow structured 1D Co@C nanotubes are synthesized by utilizing CTAB micelle as sacrificial templates. These nanotubes possess a locally conductive network formed by their tubular structure, exceptional magnetic properties confirmed through micromagnetic simulation. Experimental outcomes and RCS simulation validate that the synergistic effect of multiple loss mechanisms and optimized impedance matching contributes to the outstanding electromagnetic wave absorption capabilities.image

Spectropolarimetric observations used to infer the solar magnetic fields are obtained with a limited spatial resolution. The effects of this limited resolution on the inference of the open flux over the observed region have not been extensively studied. We aim to characterize the biases that arise in the inference of the mean flux density by performing an end-to-end study that involves the generation of synthetic data, its interpretation (inversion), and a comparison of the results with the original model. We synthesized polarized spectra of the two magnetically sensitive lines of neutral iron around 630\,nm from a state-of-the-art numerical simulation of the solar photosphere. We then performed data degradation to simulate the effect of the telescope with a limited angular resolution and interpreted (inverted) the data using a Milne-Eddington spectropolarimetric inversion code. We then studied the dependence of the inferred parameters on the telescope resolution. The results show a significant decrease in the mean magnetic flux density -- related to the open flux observed at the disk center -- with decreasing telescope resolution. The original net magnetic field flux is fully resolved by a 1m telescope, but a 20\,cm aperture telescope yields a 30\% smaller value. Even in the fully resolved case, the result is still biased due to the corrugation of the photospheric surface. Even the spatially averaged quantities, such as the open magnetic flux in the observed region, are underestimated when the magnetic structures are unresolved. The reason for this is the presence of nonlinearities in the magnetic field inference process. This effect might have implications for the modeling of large-scale solar magnetic fields; for example, those corresponding to the coronal holes, or the polar magnetic fields, which are relevant to our understanding of the solar cycle., Comment: To be published in A&A

Poor generalization performance caused by distribution shifts in unseen domains often hinders the trustworthy deployment of deep neural networks. Many domain generalization techniques address this problem by adding a domain invariant regularization loss terms during training. However, there is a lack of modular software that allows users to combine the advantages of different methods with minimal effort for reproducibility. DomainLab is a modular Python package for training user specified neural networks with composable regularization loss terms. Its decoupled design allows the separation of neural networks from regularization loss construction. Hierarchical combinations of neural networks, different domain generalization methods, and associated hyperparameters, can all be specified together with other experimental setup in a single configuration file. Hierarchical combinations of neural networks, different domain generalization methods, and associated hyperparameters, can all be specified together with other experimental setup in a single configuration file. In addition, DomainLab offers powerful benchmarking functionality to evaluate the generalization performance of neural networks in out-of-distribution data. The package supports running the specified benchmark on an HPC cluster or on a standalone machine. The package is well tested with over 95 percent coverage and well documented. From the user perspective, it is closed to modification but open to extension. The package is under the MIT license, and its source code, tutorial and documentation can be found at https://github.com/marrlab/DomainLab.

We address the online combinatorial choice of weight multipliers for multi-objective optimization of many loss terms parameterized by neural works via a probabilistic graphical model (PGM) for the joint model parameter and multiplier evolution process, with a hypervolume based likelihood promoting multi-objective descent. The corresponding parameter and multiplier estimation as a sequential decision process is then cast into an optimal control problem, where the multi-objective descent goal is dispatched hierarchically into a series of constraint optimization sub-problems. The subproblem constraint automatically adapts itself according to Pareto dominance and serves as the setpoint for the low level multiplier controller to schedule loss landscapes via output feedback of each loss term. Our method is multiplier-free and operates at the timescale of epochs, thus saves tremendous computational resources compared to full training cycle multiplier tuning. It also circumvents the excessive memory requirements and heavy computational burden of existing multi-objective deep learning methods. We applied it to domain invariant variational auto-encoding with 6 loss terms on the PACS domain generalization task, and observed robust performance across a range of controller hyperparameters, as well as different multiplier initial conditions, outperforming other multiplier scheduling methods. We offered modular implementation of our method, admitting extension to custom definition of many loss terms.

We study joint learning of network topology and a mixed opinion dynamics, in which agents may have different update rules. Such a model captures the diversity of real individual interactions. We propose a learning algorithm based on multi-armed bandit algorithms to address the problem. The goal of the algorithm is to find each agent's update rule from several candidate rules and to learn the underlying network. At each iteration, the algorithm assumes that each agent has one of the updated rules and then modifies network estimates to reduce validation error. Numerical experiments show that the proposed algorithm improves initial estimates of the network and update rules, decreases prediction error, and performs better than other methods such as sparse linear regression and Gaussian process regression., Part of ISBN 9781713872344QC 20240208

Designing Luenberger observers for nonlinear systems involves the challenging task of transforming the state to an alternate coordinate system, possibly of higher dimensions, where the system is asymptotically stable and linear up to output injection. The observer then estimates the system's state in the original coordinates by inverting the transformation map. However, finding a suitable injective transformation whose inverse can be derived remains a primary challenge for general nonlinear systems. We propose a novel approach that uses supervised physics-informed neural networks to approximate both the transformation and its inverse. Our method exhibits superior generalization capabilities to contemporary methods and demonstrates robustness to both neural network's approximation errors and system uncertainties., Part of ISBN 979-8-3503-2806-6QC 20230922

Delta ($\delta$) sunspots sometimes host fast photospheric flows along the central magnetic polarity inversion line (PIL). Here we study the strong Doppler shift signature in the central penumbral light bridge of solar active region NOAA 12673. Observations from the Helioseismic and Magnetic Imager (HMI) indicate highly sheared, strong magnetic fields. Large Doppler shifts up to 3.2 km s$^{-1}$ appeared during the formation of the light bridge and persisted for about 16 hours. A new velocity estimator, called DAVE4VMwDV, reveals fast converging and shearing motion along the PIL from HMI vector magnetograms, and recovers the observed Doppler signal much better than an old version of the algorithm. The inferred velocity vectors are largely (anti-)parallel to the inclined magnetic fields, suggesting that the observed Doppler shift contains significant contribution from the projected, field-aligned flows. High-resolution observations from the Hinode/Spectro-Polarimeter (SP) further exhibit a clear correlation between the Doppler velocity and the cosine of the magnetic inclination, which is in agreement with HMI results and consistent with a field-aligned flow of about 9.6 km s$^{-1}$. The complex Stokes profiles suggest significant gradients of physical variables along the line of sight. We discuss the implications on the $\delta$-spot magnetic structure and the flow-driving mechanism., Comment: 17 pages, 14 figures. Accepted for publication in ApJ

We study the extent to which Milne-Eddington inversions are able to retrieve and characterize the magnetic landscape of the solar poles from observations by the spectropolarimeter onboard Hinode. In particular, we evaluate whether a variable magnetic filling factor is an adequate modeling technique for retrieving the intrinsic magnetic properties from every pixel in the polar field of view. We first generate synthetic spectra emerging from a numerical simulation of a "plage" region at an inclined line of sight of 65$^{\circ}$, and degrade the data to emulate real observations. Then, we invert the synthetic spectra with two Milne-Eddington inversion codes that feature different treatments of the magnetic filling factor, and relate the retrieved magnetic quantities back to their original values in the simulation cube. We find that while the apparent retrieved magnetic properties map well the spatially-degraded simulation, the intrinsic magnetic quantities bear little relation to the magnetic field at the native resolution of the simulation. We discuss the systematic biases caused by line-of-sight foreshortening, spatial degradation, photon noise and modeling assumptions embedded in the inversion algorithm., Comment: Accepted for publication in ApJ

Using time-distance local helioseismology flow maps within 1 Mm of the solar photosphere, we detect inflows toward activity belts that contribute to solar cycle scale variations in near-surface meridional flow. These inflows stretch out as far as 30 degrees away from active region centroids. If active region neighborhoods are excluded, the solar cycle scale variation in background meridional flow diminishes to below 2~m~s$^{-1}$, but still shows systematic variations in the absence of active regions between Sunspot Cycles 24 and 25. We, therefore, propose that the near-surface meridional flow is a three component flow made up of: a constant baseline flow profile that can be derived from quiet Sun regions, variations due to inflows around active regions, and solar cycle scale variation of the order of 2~m~s$^{-1}$. Torsional oscillation, on the other hand, is found to be a global phenomenon i.e. exclusion of active region neighborhoods does not affect its magnitude or phase significantly. This non-variation of torsional oscillation with distance away from active regions and the three-component breakdown of the near-surface meridional flow serve as vital constraints for solar dynamo models and surface flux transport simulations., Comment: 14 pages, 9 figures, accepted for publication in the Astrophysical Journal

We identify 182 flares on 158 stars within 100 pc of the Sun in both the near-ultraviolet (NUV: 1750-2750 \r{A}) and far-ultraviolet (FUV: 1350-1750 \r{A}) using high-cadence light curves from the Galaxy Evolution Explorer (GALEX). Ultraviolet (UV) emission from stellar flares plays a crucial role in determining the habitability of exoplanetary systems. However, whether such UV emission promotes or threatens such life depends strongly on the energetics of these flares. Most studies assessing the effect of flares on planetary habitability assume a 9000 K blackbody spectral energy distribution that produces more NUV flux than FUV flux ($R \equiv F_{\rm FUV} / F_{\rm NUV} \approx \frac{1}{6}$). Instead, we observe the opposite with the excess FUV reaching $R \approx \frac{1}{2} - 2$, roughly $3-12$ times the expectation of a 9000 K blackbody. The ratio of FUV to NUV time-integrated flare energies is 3.0 times higher on average than would be predicted by a constant 9000 K blackbody during the flare. Finally, we find that the FUV/NUV ratio at peak tentatively correlates ($\sim 2 \sigma$ significance) both with total UV flare energy and with the G - RP color of the host star. On average, we observe higher FUV/NUV ratios at peak in $E_{\text{UV}}>10^{32}$ erg flares and in flares on fully convective stars., Comment: Submitted to MNRAS, comments welcome

M-dwarf flares observed by the \textit{Transiting Exoplanet Survey Satellite} (\textit{TESS}) sometimes exhibit a "peak-bump" light-curve morphology, characterized by a secondary, gradual peak well after the main, impulsive peak. A similar "late phase" is frequently detected in solar flares observed in the extreme-ultraviolet from longer hot coronal loops distinct from the impulsive flare structures. White-light emission has also been observed in off-limb solar flare loops. Here, we perform a suite of one-dimensional hydrodynamic loop simulations for M-dwarf flares inspired by these solar examples. Our results suggest that coronal plasma condensation following impulsive flare heating can yield high electron number density in the loop, allowing it to contribute significantly to the optical light curves via free-bound and free-free emission mechanisms. Our simulation results qualitatively agree with \textit{TESS} observations: the longer evolutionary time scale of coronal loops produces a distinct, secondary emission peak; its intensity increases with the injected flare energy. We argue that coronal plasma condensation is a possible mechanism for the \textit{TESS} late-phase flares., Comment: 31 pages, 13 figures, accepted for publication in ApJ

With the aim of investigating how the magnetic field in solar active regions (ARs) controls flare activity, i.e., whether a confined or eruptive flare occurs, we analyze 106 flares of Geostationary Operational Environmental Satellite (GOES) class $\geq$M1.0 during 2010$-$2019. We calculate mean characteristic twist parameters $\alpha$$_{FPIL}$ within the "flaring polarity inversion line" region and $\alpha$$_\mathrm{HFED}$ within the area of high photospheric magnetic free energy density, which both provide measures of the nonpotentiality of AR core region. Magnetic twist is thought to be related to the driving force of electric current-driven instabilities, such as the helical kink instability. We also calculate total unsigned magnetic flux ($\Phi$$_\mathrm{AR}$) of ARs producing the flare, which describes the strength of the background field confinement. By considering both the constraining effect of background magnetic fields and the magnetic non-potentiality of ARs, we propose a new parameter $\alpha$/$\Phi$$_\mathrm{AR}$ to measure the probability for a large flare to be associated with a coronal mass ejection (CME). We find that in about 90\% of eruptive flares, $\alpha$$_\mathrm{FPIL}$/$\Phi$$_\mathrm{AR}$ and $\alpha$$_\mathrm{HFED}$/$\Phi$$_\mathrm{AR}$ are beyond critical values (2.2$\times$$10^{-24}$ and 3.2$\times$$10^{-24}$ Mm$^{-1}$ Mx$^{-1}$), whereas they are less than critical values in $\sim$ 80\% of confined flares. This indicates that the new parameter $\alpha$/$\Phi$$_\mathrm{AR}$ is well able to distinguish eruptive flares from confined flares. Our investigation suggests that the relative measure of magnetic nonpotentiality within the AR core over the restriction of the background field largely controls the capability of ARs to produce eruptive flares., Comment: 7 pages, 4 figures, 1 table and 1 online database

Ag/Y2 O3 has excellent potential to replace Ag/CdO as the environmentally friendly electrical contact material. Using spherical Y2 O3 as the starting material, Ag/Y2 O3 contacts with a quasi-continuous network structure were successfully fabricated by a low-energy ball milling treatment. The mean size of Y2 O3 used ranged from 243 to 980 nm. Due to the differences in the size of Y2 O3, Ag/Y2 O3 contacts had different primitive microstructures, thereby exhibiting distinctive anti-arc-erosion capabilities. Ag/Y2 O3 contact prepared using 243 nm Y2 O3 showed the best anti-arc-erosion capability and the most outstanding electrical performance measures, such as low contact resistance, less mass transfer, and no failure up to 105 cycle times. The quasi-continuous network structure formed in the micro-scale was responsible for the excellent electrical performance. The short distance between Y2 O3 particles in the network promoted the cathode arc motion, and thus alleviated the localized erosion. The results obtained herein may inspire further attempts to design electrical contacts rationally. © 2022 by the authors. Licensee MDPI, Basel, Switzerland.

Designing Luenberger observers for nonlinear systems involves the challenging task of transforming the state to an alternate coordinate system, possibly of higher dimensions, where the system is asymptotically stable and linear up to output injection. The observer then estimates the system's state in the original coordinates by inverting the transformation map. However, finding a suitable injective transformation whose inverse can be derived remains a primary challenge for general nonlinear systems. We propose a novel approach that uses supervised physics-informed neural networks to approximate both the transformation and its inverse. Our method exhibits superior generalization capabilities to contemporary methods and demonstrates robustness to both neural network's approximation errors and system uncertainties., Comment: Proceedings of the 2023 American Control Conference (ACC)

Despite investments in multiple space and ground-based solar observatories by the global community, the Sun's polar regions remain unchartered territory - the last great frontier for solar observations. Breaching this frontier is fundamental to understanding the solar cycle - the ultimate driver of short-to-long term solar activity that encompasses space weather and space climate. Magnetohydrodynamic dynamo models and empirically observed relationships have established that the polar field is the primary determinant of the future solar cycle amplitude. Models of solar surface evolution of tilted active regions indicate that the mid to high latitude surges of magnetic flux govern dynamics leading to the reversal and build-up of polar fields. Our theoretical understanding and numerical models of this high latitude magnetic field dynamics and plasma flows - that are a critical component of the sunspot cycle - lack precise observational constraints. This limitation compromises our ability to observe the enigmatic kilo Gauss polar flux patches and constrain the polar field distribution at high latitudes. The lack of these observations handicap our understanding of how high latitude magnetic fields power polar jets, plumes, and the fast solar wind that extend to the boundaries of the heliosphere and modulate solar open flux and cosmic ray flux within the solar system. Accurate observation of the Sun's polar regions, therefore, is the single most outstanding challenge that confronts Heliophysics. This paper argues the scientific case for novel out of ecliptic observations of the Sun's polar regions, in conjunction with existing, or future multi-vantage point heliospheric observatories. Such a mission concept can revolutionize the field of Heliophysics like no other mission concept has - with relevance that transcends spatial regimes from the solar interior to the heliosphere., Comment: This White Paper was submitted in 2022 to the United States National Academies Solar and Space Physics (Heliophysics) Decadal Survey

Rapid increase of horizontal magnetic field ($B_h$) around the flaring polarity inversion line is the most prominent photospheric field change during flares. It is considered to be caused by the contraction of flare loops, the details behind which is still not fully understood. Here we investigate the $B_h$-increase in 35 major flares using HMI high-cadence vector magnetograms. We find that $B_h$-increase is always accompanied by the increase of field inclination. It usually initiates near the flare ribbons, showing step-like change in between the ribbons. In particular, its evolution in early flare phase shows close spatio-temporal correlation to flare ribbons. We further find that $B_h$-increase tends to have similar intensity in confined and eruptive flares, but larger spatial-extent in eruptive flares in a statistical sense. Its intensity and timescale have inverse and positive correlations to the initial ribbon separations, respectively. The results altogether are well consistent with a recent proposed scenario which suggests that the reconnection-driven contraction of flare loops enhances photospheric $B_h$ according to the ideal induction equation, providing statistical evidence to the reconnection-driven origin for $B_h$-increase for the first time., Comment: 14 pages, 5 figures, 2 tables. Accepted for publication in ApJL

The development of a simple and general approach for manufacturing complex hollow structures is essential to achieve rapid dynamic response and high energy density storage for supercapacitor electrodes. Herein, we report a versatile heating rate-controlled synthetic route to prepare MnCo2O4 microspheres with multiple interiors. Comparative study reveals that MnCo2O4 microspheres with core–shell structure (MCO-CSS) process the highest specific surface area (91.6 m2 g−1) and oxygen vacancy density (0.16), which can provide more electroactive sites for surface Faradic reaction and further facilitate the kinetics of rapid charge storage. While the porous shell and well-defined gap endow MCO-CSS with excellent permeability and low-resistance pathways for electrolyte ions, as well as enhanced volume change accommodation. Benefiting from the structural advantages, MCO-CSS displays an excellent specific capacitance of 1064.4F g−1 at 1 A g−1, remarkable rate capability, and satisfactory cycle stability (89.1% capacitance retention at 2 A g−1 over 1000 cycles). Additionally, an asymmetric supercapacitor device, with an MCO-CSS cathode and porous carbon spheres anode, delivers a remarkable energy density of 33.9 Wh kg−1 at a power density of 187.4 W kg−1, as well as 84.13% capacitance retention over 3000 cycles at 1 A g−1. © 2022 Elsevier B.V.

Layered double hydroxides (LDHs) have attracted broad attention as cathode materials for hybrid supercapacitors (HSCs) because of their ultrahigh theoretical specific capacitance, high compositional flexibility, and adjustable interlayer spacing. However, as reported, specific capacitance of LDHs is still far below the theoretical value, inspiring countless efforts to these ongoing challenges. Herein, a hierarchical nanocage structure assembled by NiCo-LDH nanosheet arrays was rationally designed and fabricated via a facile solvothermal method assisted by the ZIF-67 template. The transformation from the ZIF-67 template to this hollow structure is achieved by a synergistic effect involving the Kirkendall effect and the Ostwald ripening process. The enlarged specific surface area co-occurred with broadened interlayer spacing of LDH nanosheets by finely increasing the Ni concentration, leading to synchronous improvement of electron/ion transfer kinetics. The optimized NiCo-LDH-210 electrode displays a maximum specific capacitance of 2203.6 F g-1 at 2 A g-1, excellent rate capability, and satisfactory cycling stability because of the highly exposed active sites and shortened ion transport paths provided by vertically aligned LDH nanosheets together with the cavity. Furthermore, the assembled HSC device achieves a superior energy density of 57.3 Wh kg-1 with prominent cycling stability. Impressively, the design concept of complex construction derived from metal-organic frameworks (MOF) derivatives shows tremendous potential for use in energy storage systems. © 2022 American Chemical Society.

Lateral superlattices have been a research focus due to the strain induced in their coherent structure, while the growth of a large size heterostructure with high density of heterointerfaces remains challenging. In this work, we report a gap-filling approach to synthesize large-scale lateral mesh heterostructures of WS2 embedded in a MoS2 matrix. This synthesis method utilizes the uniformly distributed gaps in single crystalline MoS2, made by the sulfur substitution-induced transformation of metastable-MoTe2 as the growth pattern, for the second material growth. By finely controlling the growth kinetics, a highly crystalline WS2/MoS2 lateral heterostructure mesh is successfully grown. Optical images and Raman mapping show a clear spatial distribution of WS2 channels embedded in MoS2. The coherent epitaxial nature is further confirmed by scanning transmission electron microscopy, showing insignificant dislocation at the interfaces. Density function theory simulation suggests that the growth of WS2 starts from the edges of cracked MoS2. The MoS2/WS2/MoS2 heterostructure device demonstrates an intrinsic electric barrier formed at the interfaces. This work provides a new tool for the practical preparation of large-scale high density 2D heterostructures. © 2022 American Chemical Society.

Zixing Bai1 *, Xuhan Cao1 *, Yanjun Yang,1 Xudong Sun,1 Yongli Dong,2 Jianmin Wen,1 Weidong Sun1 1Second Department of Orthopedics, Wangjing Hospital of China Academy of Chinese Medical Sciences, Beijing, People’s Republic of China; 2Scientific Research Department, Wangjing Hospital of China Academy of Chinese Medical Sciences, Beijing, People’s Republic of China*These authors contributed equally to this workCorrespondence: Weidong Sun, Wangjing Hospital of China Academy of Chinese Medical Sciences, No. 6 Central South Road, Wangjing, Chaoyang District, Beijing, 100102, People’s Republic of China, Tel +86-84739140, Email sunweidong8239@aliyun.comObjective: To investigate the risk factors for hallux valgus complicated with pain under the second metatarsal and construct an effective model and method for predicting hallux valgus complicated with pain under the second metatarsal based on risk factors.Methods: A total of 545 patients with hallux valgus who were admitted to our hospital were divided randomly into a training set and a validation set. The demographic characteristics, imaging indices and gait test indices of the patients were collected. The risk factors were identified by univariate and multivariate logistic regression analyses. A risk prediction model for hallux valgus with pain under the second metatarsal was established, and the area under the curve (AUC) of the receiver operating characteristic and a decision curve analysis were used for verification and identification. The value of the model was tested in the verification group.Results: Second metatarsal length, second metatarsal peak pressure, hallux valgus angle (HVA), intermetatarsal angle 1– 2 (IMA1– 2) and weight were the risk factors for hallux valgus complicated with pain under the second metatarsal. Based on the weighting of these seven risk factors, a prediction model was established. The AUC of the prediction model was 0.84 (95% confidence interval [CI]: 0.802

Bald patch (BP) is a magnetic topological feature where U-shaped field lines turn tangent to the photosphere. Field lines threading the BP trace a separatrix surface where reconnection preferentially occurs. Here we study the evolution of multiple, strong-field BPs in active region 12673 during the most intense, X9.3 flare of solar cycle 24. The central BP, located between the initial flare ribbons, largely "disintegrated" within 35 minutes. The more remote, southern BP survived. The disintegration manifested as a 9-degree rotation of the median shear angle; the perpendicular component of the horizontal field (with respect to the polarity inversion line) changed sign. The parallel component exhibited a step-wise, permanent increase of 1 kG, consistent with previous observations of the flare-related "magnetic imprint". The observations suggest that magnetic reconnection during a major eruption may involve entire BP separatrices, leading to a change of magnetic topology from BPs to sheared arcades., Comment: 7 pages, 5 figures, 2 animations via Zenodo. Accepted for publication in ApJL

Whilst intense solar flares are almost always accompanied by a coronal mass ejection (CME), reports on stellar CMEs are rare, despite the frequent detection of stellar 'super flares'. The torus instability of magnetic flux ropes is believed to be one of the main driving mechanisms of solar CMEs. Suppression of the torus instability, due to a confining background coronal magnetic field that decreases sufficiently slowly with height, may contribute to the lack of stellar CME detection. Here we use the solar magnetic field as a template to estimate the vertical extent of this 'torus-stable zone' (TSZ) above a stellar active region. For an idealised potential field model comprising the fields of a local bipole (mimicking a pair of starspots) and a global dipole, we show that the upper bound of the TSZ increases with the bipole size, the dipole strength, and the source surface radius where the coronal field becomes radial. The boundaries of the TSZ depend on the interplay between the spots' and the dipole's magnetic fields, which provide the local- and global-scale confinement, respectively. They range from about half the bipole size to a significant fraction of the stellar radius. For smaller spots and an intermediate dipole field, a secondary TSZ arises at a higher altitude, which may increase the likelihood of 'failed eruptions'. Our results suggest that the low apparent CME occurrence rate on cool stars is, at least partially, due to the presence of extended TSZs., Comment: 12 pages, 10 figures. Data and 3 interactive figures available via Zenodo. Accepted to MNRAS

Violent solar flares and coronal mass ejections (CMEs) are magnetic phenomena. However, how magnetic fields reconnecting in the flare differ from non-flaring magnetic fields remains unclear owing to the lack of studies of the flare magnetic properties. Here we present a first statistical study of flaring (highlighted by flare-ribbons) vector magnetic fields in the photosphere. Our systematic approach allows us to describe key physical properties of solar flare magnetism, including distributions of magnetic flux, magnetic shear, vertical current and net current over flaring versus non-flaring parts of the active region, and compare these with flare/CME properties. Our analysis suggests that while flares are guided by the physical properties that scale with AR size, like the total amount of magnetic flux that participates in the reconnection process and the total current (extensive properties), CMEs are guided by mean properties, like the fraction of the AR magnetic flux that participates (intensive property), with little dependence on the amount of shear at polarity inversion line (PIL) or the net current. We find that the non-neutralized current is proportional to the amount of shear at PIL, providing direct evidence that net vertical currents are formed as a result of any mechanism that could generate magnetic shear along PIL. We also find that eruptive events tend to have smaller PIL fluxes and larger magnetic shears than confined events. Our analysis provides a reference for more realistic solar and stellar flare models. The database is available online and can be used for future quantitative studies of flare magnetism., Comment: 25 pages, 12 figures, accepted to The Astrophysical Journal (10 November 2021); Link to FlareMagDB dataset: http://solarmuri.ssl.berkeley.edu/~kazachenko/FlareMagDB

We present a new data product, called Space-Weather MDI Active Region Patches (SMARPs), derived from maps of the solar surface magnetic field taken by the Michelson Doppler Imager (MDI) aboard the Solar and Heliospheric Observatory (SoHO). Together with the Space-Weather HMI Active Region Patches (SHARPs), derived from similar maps taken by the Helioseismic and Magnetic Imager (HMI) aboard the Solar Dynamics Observatory, these data provide a continuous and seamless set of maps and keywords that describe every active region observed over the last two solar cycles, from 1996 to the present day. In this paper, we describe the SMARP data and compare it to the SHARP data., Comment: 11 pages, 3 figures, accepted in ApJS

We investigate the orientation of the photospheric magnetic fields in the solar polar region using observations from the Helioseismic and Magnetic Imager (HMI). Inside small patches of significant polarization, the inferred magnetic field vectors at $1''$ scale appear to systematically deviate from the radial direction. Most tilt towards the pole; all are more inclined toward the plane of sky compared to the radial vector. These results, however, depend on the "filling factor" $f$ that characterizes the unresolved magnetic structures. The default, uninformative $f\equiv1$ for HMI will incur larger inclination and less radial fields than $f<1$. The observed trend may be a systematic bias inherent to the limited resolution., Comment: 3 pages, 1 figure. Accepted for publication in RNAAS

Dark matter can decay via gravity portals and contribute a minor decay (MD) width. This paper proposes that nonminimal coupling between the Higgs field and gravity can lead to additional gravity-Higgs portals for gravitational dark matter to decay and contribute additional decay (AD) width. For scalar dark matter candidates with a mass less than 5 TeV, the decay branching ratios are almost independent of whether dark matter decay through gravity portals or gravity-Higgs portals; therefore, it is almost impossible to distinguish gravity portals from gravity-Higgs portals by observing decay products. For fermionic singlet dark matter candidates, their MD products are diverse, while their AD products are monotonous, only neutrinos and Higgs bosons. As for both dark matter candidates, gravity-Higgs portals prefer to dominate the decay at around several hundred GeV. Besides, for both TeV dark matter candidates, there is still a vast parameter space alive.

Emerging dimming occurs in isolated solar active regions (ARs) during the early stages of magnetic flux emergence. Observed by the Atmospheric Imaging Assembly, it features a rapid decrease in extreme-ultraviolet (EUV) emission in the 171 \r{A} channel images, and a simultaneous increase in the 211 \r{A} images. Here, we analyze the coronal thermodynamic and magnetic properties to probe its physical origin. We calculate the time-dependent differential emission measures for a sample of 18 events between 2010 and 2012. The emission measure (EM) decrease in the temperature range $5.7 \le \log_{10}T \le 5.9$ is well correlated with the EM increase in $6.2 \le \log_{10}T \le 6.4$ over eight orders of magnitude. This suggests that the coronal plasma is being heated from the quiet-Sun, sub-MK temperature to 1-2 MK, more typical for ARs. Potential field extrapolation indicates significant change in the local magnetic connectivity: the dimming region is now linked to the newly emerged flux via longer loops. We conclude that emerging dimming is likely caused by coronal heating episodes, powered by reconnection between the emerging and the ambient magnetic fields., Comment: 8 pages, 7 figures, 1 table. Accepted for publication in the Astrophysical Journal

High-energy radiation from the Sun governs the behavior of Earth's upper atmosphere and such radiation from any planet-hosting star can drive the long-term evolution of a planetary atmosphere. However, much of this radiation is unobservable because of absorption by Earth's atmosphere and the interstellar medium. This motivates the identification of a proxy that can be readily observed from the ground. Here, we evaluate absorption in the near-infrared 1083 nm triplet line of neutral orthohelium as a proxy for extreme ultraviolet (EUV) emission in the 30.4 nm line of He II and 17.1 nm line of Fe IX from the Sun. We apply deep learning to model the non-linear relationships, training and validating the model on historical, contemporaneous images of the solar disk acquired in the triplet He I line by the ground-based SOLIS observatory and in the EUV by the NASA Solar Dynamics Observatory. The model is a fully-convolutional neural network (FCN) that incorporates spatial information and accounts for the projection of the spherical Sun to 2-d images. Using normalized target values, results indicate a median pixel-wise relative error of 20% and a mean disk-integrated flux error of 7% on a held-out test set. Qualitatively, the model learns the complex spatial correlations between He I absorption and EUV emission has a predictive ability superior to that of a pixel-by-pixel model; it can also distinguish active regions from high-absorption filaments that do not result in EUV emission., Comment: Accepted to ApJ Letters

We address the task of domain generalization, where the goal is to train a predictive model such that it is able to generalize to a new, previously unseen domain. We choose a hierarchical generative approach within the framework of variational autoencoders and propose a domain-unsupervised algorithm that is able to generalize to new domains without domain supervision. We show that our method is able to learn representations that disentangle domain-specific information from class-label specific information even in complex settings where domain structure is not observed during training. Our interpretable method outperforms previously proposed generative algorithms for domain generalization as well as other non-generative state-of-the-art approaches in several hierarchical domain settings including sequential overlapped near continuous domain shift. It also achieves competitive performance on the standard domain generalization benchmark dataset PACS compared to state-of-the-art approaches which rely on observing domain-specific information during training, as well as another domain unsupervised method. Additionally, we proposed model selection purely based on Evidence Lower Bound (ELBO) and also proposed weak domain supervision where implicit domain information can be added into the algorithm., Comment: Presented at ICLR 2021 RobustML Workshop

Jianhua Chen,1,* Quanhe Chen,2,* Feilong Zhang,1 Xuehai Chen,1 Zhe Xu,1 Qiong Jiang,1 Xudong Sun,1 Jinguo Li,1 Lianglong Chen,1 Weiwei Wang1 1Department of Cardiology, Fujian Medical University Union Hospital & Fujian Provincial Institute of Coronary Disease, Fuzhou, People’s Republic of China; 2Union Clinic Medical College, Fujian Medical University, Fuzhou, People’s Republic of China*These authors contributed equally to this workCorrespondence: Weiwei WangDepartment of Cardiology, Fujian Medical University Union Hospital & Fujian Provincial Institute of Coronary Disease, No. 29 of Xinquan Road, Gulou District, Fuzhou, 350001, People’s Republic of ChinaTel +86 591 86218022Fax +86 591 83322156Email wang_wwedh@163.comObjective: This study aimed to investigate whether the enhanced endpoint of pulmonary vein isolation (PVI; intravenous injection of adenosine-triphosphate [ATP] + pacing capture + supplemental ablation) after initial PVI can reduce the long-term recurrence rate of atrial fibrillation (AF) after PVI.Methods: Patients with paroxysmal or persistent AF undergoing catheter ablation treatment were enrolled in this study and divided into three groups according to the surgical endpoint: (1) group 1 (n = 92), in which patients were observed for 30 minutes after the initial PVI and pulmonary vein–left atrium (PV–LA) electrical conduction had not recovered; (2) group 2 (n = 99), in which patients were observed for 30 minutes after the initial PVI, then intravenously injected with ATP, and PV–LA electrical conduction had not recovered; and (3) group 3 (n = 102), in which patients were observed for 30 minutes after the initial PVI, then intravenously injected with ATP + treated with ablation line pacing, and the atrium could not be captured.Results: Patients were followed up for 12 months after the operation. Twenty-eight patients in group 1 (30.4%), 19 patients in group 2 (19.2%), and 10 pat

The Coronal Global Evolutionary Model (CGEM) provides data-driven simulations of the magnetic field in the solar corona to better understand the build-up of magnetic energy that leads to eruptive events. The CGEM project has developed six capabilities. CGEM modules (1) prepare time series of full-disk vector magnetic field observations to (2) derive the changing electric field in the solar photosphere over active-region scales. This local electric field is (3) incorporated into a surface flux transport model that reconstructs a global electric field that evolves magnetic flux in a consistent way. These electric fields drive a (4) 3D spherical magneto-frictional (SMF) model, either at high-resolution over a restricted range of solid angle or at lower resolution over a global domain, to determine the magnetic field and current density in the low corona. An SMF-generated initial field above an active region and the evolving electric field at the photosphere are used to drive (5) detailed magneto-hydrodynamic (MHD) simulations of active regions in the low corona. SMF or MHD solutions are then used to compute emissivity proxies that can be compared with coronal observations. Finally, a lower-resolution SMF magnetic field is used to initialize (6) a global MHD model that is driven by an SMF electric-field time series to simulate the outer corona and heliosphere, ultimately connecting Sun to Earth. As a demonstration, this report features results of CGEM applied to observations of the evolution of NOAA Active Region 11158 in February 2011., Comment: 19 pages, 8 figures

It is well established that magnetic free energy associated with electric currents powers solar flares and coronal mass ejections (CMEs) from solar active regions (ARs). However, the conditions that determine whether an AR will produce an eruption are not well understood. Previous work suggests that the degree to which the driving electric currents, or the sum of all currents within a single magnetic polarity, are neutralized may serve as a good proxy for assessing the ability of ARs to produce eruptions. Here, we investigate the relationship between current neutralization and flare/CME production using a sample of 15 flare-active and 15 flare-quiet ARs. All flare-quiet and 4 flare-active ARs are also CME-quiet. We additionally test the relation of current neutralization to the degree of shear along polarity inversion lines (PILs) in an AR. We find that flare-productive ARs are more likely to exhibit non-neutralized currents, specifically those that also produce a CME. We find that flare/CME-active ARs also exhibit higher degrees of PIL shear than flare/CME-quiet ARs. We additionally observe that currents become more neutralized during magnetic flux emergence in flare-quiet ARs. Our investigation suggests that current neutralization in ARs is indicative of their eruptive potential., Comment: 10 pages, 8 figures, Accepted for publication in the Astrophysical Journal, A video description can be seen here https://youtu.be/kdwzOElqmqc

Active regions (ARs) often possess an observed net electric current in a single magnetic polarity. We show that such "non-neutralized" currents can arise from a geometric projection effect when a twisted flux tube obliquely intersects the photosphere. To this end, we emulate surface maps of an emerging AR by sampling horizontal slices of a semi-torus flux tube at various heights. Although the tube has no net toroidal current, its poloidal current, when projected along the vertical direction, amounts to a significant non-neutralized component on the surface. If the tube emerges only partially as in realistic settings, the non-neutralized current will 1) develop as double ribbons near the sheared polarity inversion line, (2) positively correlate with the twist, and 3) reach its maximum before the magnetic flux. The projection effect may be important to the photospheric current distribution, in particular during the early stages of flux emergence., Comment: 19 pages, 7 figures. Accepted for publication in Solar Physics. Minor updates from v1

The Daniel K. Inouye Solar Telescope (DKIST) will revolutionize our ability to measure, understand and model the basic physical processes that control the structure and dynamics of the Sun and its atmosphere. The first-light DKIST images, released publicly on 29 January 2020, only hint at the extraordinary capabilities which will accompany full commissioning of the five facility instruments. With this Critical Science Plan (CSP) we attempt to anticipate some of what those capabilities will enable, providing a snapshot of some of the scientific pursuits that the Daniel K. Inouye Solar Telescope hopes to engage as start-of-operations nears. The work builds on the combined contributions of the DKIST Science Working Group (SWG) and CSP Community members, who generously shared their experiences, plans, knowledge and dreams. Discussion is primarily focused on those issues to which DKIST will uniquely contribute.

Although many achievements have been made since Google threw out the paradigm of federated learning (FL), there still exists much room for researchers to optimize its efficiency. In this paper, we propose a high efficient FL method equipped with the double head design aiming for personalization optimization over non-IID dataset, and the gradual model sharing design for communication saving. Experimental results show that, our method has more stable accuracy performance and better communication efficient across various data distributions than other state of art methods (SOTAs), makes it more industry-friendly.

Increasing demands for optical anticounterfeiting technology require the development of versatile luminescent materials with tunable photoluminescence properties. Herein, a number of fluorescent carbon- and oxygen-doped hexagonal boron nitride (denoted as BCNO) phosphors are found to offer a such high-tech anticounterfeiting solution. These multicolor BCNO powders, developed in a two-step process with controlled annealing and oxidation, feature rod-like particle shape, with varied luminescence properties. Studies of the optical properties of BCNO, along with other characterization, provide insight into this underexplored material. Anticounterfeiting applications are demonstrated with printed patterns which are indistinguishable to the naked eye under visible light but become highly discernible under UV irradiation. The fabricated patterns are demonstrated to be both chemically stable in corrosive environments and physically robust in mechanical bending testing. These properties render BCNO as promising and versatile anticounterfeiting material a wide variety of environments.

In this study, hexagonal boron nitride nano-sheets (h-BNNS) have been grown on polycrystalline silversubstrates via chemical vapor deposition (CVD) usingammonia borane as a precursor. The h-BNNS are of few-atomic-layer thickness and form continuous coverage over thewhole Ag substrate. The atomically thin coating posesnegligible interference to the reflectivity in the UV−visiblerange. The nanosheet coating also proves very effective inprotecting Ag foil chemically. In contrast to bare Ag foil, thecoated foil displayed only minor decolorization under highconcentration of H2S. The study indicates that h-BNNS canbe a promising protective coating for Ag based items such as jewelry or mirrors used in astronomical telescopes.

We describe the PDFI_SS software library, which is designed to find the electric field at the Sun's photosphere from a sequence of vector magnetogram and Doppler velocity measurements, and estimates of horizontal velocities obtained from local correlation tracking using the recently upgraded FLCT code. The library, a collection of Fortran subroutines, uses the "PDFI" technique described by Kazachenko et al. (2014), but modified for use in spherical, Plate-Carr\'ee geometry on a staggered grid. The domain over which solutions are found is a subset of the global spherical surface, defined by user-specified limits of colatitude and longitude. Our staggered-grid approach, based on that of Yee (1966), is more conservative and self-consistent compared to the centered, Cartesian grid used by Kazachenko et al. (2014). The library can be used to compute an end-to-end solution for electric fields from data taken by the HMI instrument aboard NASA's SDO Mission. This capability has been incorporated into the HMI pipeline processing system operating at SDO's JSOC. The library is written in a general and modular way so that the calculations can be customized to modify or delete electric field contributions, or used with other data sets. Other applications include "nudging" numerical models of the solar atmosphere to facilitate assimilative simulations. The library includes an ability to compute "global" (whole-Sun) electric field solutions. The library also includes an ability to compute Potential Magnetic Field solutions in spherical coordinates. This distribution includes a number of test programs which allow the user to test the software., Comment: 68 pages, 28 figures, 3 tables. This version (arXiv) will be updated when major changes to the software occur

Due to the sparsity and irregularity of the point cloud data, methods that directly consume points have become popular. Among all point-based models, graph convolutional networks (GCN) lead to notable performance by fully preserving the data granularity and exploiting point interrelation. However, point-based networks spend a significant amount of time on data structuring (e.g., Farthest Point Sampling (FPS) and neighbor points querying), which limit the speed and scalability. In this paper, we present a method, named Grid-GCN, for fast and scalable point cloud learning. Grid-GCN uses a novel data structuring strategy, Coverage-Aware Grid Query (CAGQ). By leveraging the efficiency of grid space, CAGQ improves spatial coverage while reducing the theoretical time complexity. Compared with popular sampling methods such as Farthest Point Sampling (FPS) and Ball Query, CAGQ achieves up to 50X speed-up. With a Grid Context Aggregation (GCA) module, Grid-GCN achieves state-of-the-art performance on major point cloud classification and segmentation benchmarks with significantly faster runtime than previous studies. Remarkably, Grid-GCN achieves the inference speed of 50fps on ScanNet using 81920 points per scene as input.

Time series classification problems have drawn increasing attention in the machine learning and statistical community. Closely related is the field of functional data analysis (FDA): it refers to the range of problems that deal with the analysis of data that is continuously indexed over some domain. While often employing different methods, both fields strive to answer similar questions, a common example being classification or regression problems with functional covariates. We study methods from functional data analysis, such as functional generalized additive models, as well as functionality to concatenate (functional-) feature extraction or basis representations with traditional machine learning algorithms like support vector machines or classification trees. In order to assess the methods and implementations, we run a benchmark on a wide variety of representative (time series) data sets, with in-depth analysis of empirical results, and strive to provide a reference ranking for which method(s) to use for non-expert practitioners. Additionally, we provide a software framework in R for functional data analysis for supervised learning, including machine learning and more linear approaches from statistics. This allows convenient access, and in connection with the machine-learning toolbox mlr, those methods can now also be tuned and benchmarked.

Aiming at a comprehensive and concise tutorial survey, recap of variational inference and reinforcement learning with Probabilistic Graphical Models are given with detailed derivations. Reviews and comparisons on recent advances in deep reinforcement learning are made from various aspects. We offer detailed derivations to a taxonomy of Probabilistic Graphical Model and Variational Inference methods in deep reinforcement learning, which serves as a complementary material on top of the original contributions., Comment: 2019 IEEE Symposium on Computational Intelligence, Symposium on Adaptive Dynamic Programming and Reinforcement Learning

A novel variational inference based resampling framework is proposed to evaluate the robustness and generalization capability of deep learning models with respect to distribution shift. We use Auto Encoding Variational Bayes to find a latent representation of the data, on which a Variational Gaussian Mixture Model is applied to deliberately create distribution shift by dividing the dataset into different clusters. Wasserstein distance is used to characterize the extent of distribution shift between the generated data splits. We compare several popular Convolutional Neural Network (CNN) architectures and Bayesian CNN models for image classification on the Fashion-MNIST dataset, to assess their robustness and generalization behavior under the deliberately created distribution shift, as well as under random Cross Validation. Our method of creating artificial domain splits of a single dataset can also be used to establish novel model selection criteria and assessment tools in machine learning, as well as benchmark methods for domain adaptation and domain generalization approaches.

In Multi-Goal Reinforcement Learning, an agent learns to achieve multiple goals with a goal-conditioned policy. During learning, the agent first collects the trajectories into a replay buffer, and later these trajectories are selected randomly for replay. However, the achieved goals in the replay buffer are often biased towards the behavior policies. From a Bayesian perspective, when there is no prior knowledge about the target goal distribution, the agent should learn uniformly from diverse achieved goals. Therefore, we first propose a novel multi-goal RL objective based on weighted entropy. This objective encourages the agent to maximize the expected return, as well as to achieve more diverse goals. Secondly, we developed a maximum entropy-based prioritization framework to optimize the proposed objective. For evaluation of this framework, we combine it with Deep Deterministic Policy Gradient, both with or without Hindsight Experience Replay. On a set of multi-goal robotic tasks of OpenAI Gym, we compare our method with other baselines and show promising improvements in both performance and sample-efficiency., Comment: Published in International Conference on Machine Learning (ICML 2019), Long Beach, USA. arXiv admin note: text overlap with arXiv:1902.08039

Machine learning pipeline potentially consists of several stages of operations like data preprocessing, feature engineering and machine learning model training. Each operation has a set of hyper-parameters, which can become irrelevant for the pipeline when the operation is not selected. This gives rise to a hierarchical conditional hyper-parameter space. To optimize this mixed continuous and discrete conditional hierarchical hyper-parameter space, we propose an efficient pipeline search and configuration algorithm which combines the power of Reinforcement Learning and Bayesian Optimization. Empirical results show that our method performs favorably compared to state of the art methods like Auto-sklearn , TPOT, Tree Parzen Window, and Random Search.

A novel machine learning optimization process coined Restrictive Federated Model Selection (RFMS) is proposed under the scenario, for example, when data from healthcare units can not leave the site it is situated on and it is forbidden to carry out training algorithms on remote data sites due to either technical or privacy and trust concerns. To carry out a clinical research under this scenario, an analyst could train a machine learning model only on local data site, but it is still possible to execute a statistical query at a certain cost in the form of sending a machine learning model to some of the remote data sites and get the performance measures as feedback, maybe due to prediction being usually much cheaper. Compared to federated learning, which is optimizing the model parameters directly by carrying out training across all data sites, RFMS trains model parameters only on one local data site but optimizes hyper-parameters across other data sites jointly since hyper-parameters play an important role in machine learning performance. The aim is to get a Pareto optimal model with respective to both local and remote unseen prediction losses, which could generalize well across data sites. In this work, we specifically consider high dimensional data with shifted distributions over data sites. As an initial investigation, Bayesian Optimization especially multi-objective Bayesian Optimization is used to guide an adaptive hyper-parameter optimization process to select models under the RFMS scenario. Empirical results show that solely using the local data site to tune hyper-parameters generalizes poorly across data sites, compared to methods that utilize the local and remote performances. Furthermore, in terms of dominated hypervolumes, multi-objective Bayesian Optimization algorithms show increased performance across multiple data sites among other candidates.

It is widely believed that loops observed in the solar atmosphere trace out magnetic field lines. However, the degree to which magnetic field extrapolations yield field lines that actually do follow loops has yet to be studied systematically. In this paper we apply three different extrapolation techniques - a simple potential model, a NLFF model based on photospheric vector data, and a NLFF model based on forward fitting magnetic sources with vertical currents - to 15 active regions that span a wide range of magnetic conditions. We use a distance metric to assess how well each of these models is able to match field lines to the 12,202 loops traced in coronal images. These distances are typically 1-2". We also compute the misalignment angle between each traced loop and the local magnetic field vector, and find values of 5-12$^\circ$. We find that the NLFF models generally outperform the potential extrapolation on these metrics, although the differences between the different extrapolations are relatively small. The methodology that we employ for this study suggests a number of ways that both the extrapolations and loop identification can be improved., Comment: Accepted for publication in ApJ