Search

Recent advances in machine learning have highlighted Federated Learning (FL) as a promising approach that enables multiple distributed users (so-called clients) to collectively train ML models without sharing their private data. While this privacy-preserving method shows potential, it struggles when data across clients is not independent and identically distributed (non-IID) data. The latter remains an unsolved challenge that can result in poorer model performance and slower training times. Despite the significance of non-IID data in FL, there is a lack of consensus among researchers about its classification and quantification. This technical survey aims to fill that gap by providing a detailed taxonomy for non-IID data, partition protocols, and metrics to quantify data heterogeneity. Additionally, we describe popular solutions to address non-IID data and standardized frameworks employed in FL with heterogeneous data. Based on our state-of-the-art survey, we present key lessons learned and suggest promising future research directions.

Training and deploying Machine Learning models that simultaneously adhere to principles of fairness and privacy while ensuring good utility poses a significant challenge. The interplay between these three factors of trustworthiness is frequently underestimated and remains insufficiently explored. Consequently, many efforts focus on ensuring only two of these factors, neglecting one in the process. The decentralization of the datasets and the variations in distributions among the clients exacerbate the complexity of achieving this ethical trade-off in the context of Federated Learning (FL). For the first time in FL literature, we address these three factors of trustworthiness. We introduce PUFFLE, a high-level parameterised approach that can help in the exploration of the balance between utility, privacy, and fairness in FL scenarios. We prove that PUFFLE can be effective across diverse datasets, models, and data distributions, reducing the model unfairness up to 75%, with a maximum reduction in the utility of 17% in the worst-case scenario, while maintaining strict privacy guarantees during the FL training.

As Automatic Speech Recognition (ASR) models become ever more pervasive, it is important to ensure that they make reliable predictions under corruptions present in the physical and digital world. We propose Speech Robust Bench (SRB), a comprehensive benchmark for evaluating the robustness of ASR models to diverse corruptions. SRB is composed of 114 input perturbations which simulate an heterogeneous range of corruptions that ASR models may encounter when deployed in the wild. We use SRB to evaluate the robustness of several state-of-the-art ASR models and observe that model size and certain modeling choices such as the use of discrete representations, or self-training appear to be conducive to robustness. We extend this analysis to measure the robustness of ASR models on data from various demographic subgroups, namely English and Spanish speakers, and males and females. Our results revealed noticeable disparities in the model's robustness across subgroups. We believe that SRB will significantly facilitate future research towards robust ASR models, by making it easier to conduct comprehensive and comparable robustness evaluations., Comment: submitted to NeurIPS datasets and benchmark track 2025

AT 2019azh is a H+He tidal disruption event (TDE) with one of the most extensive ultraviolet and optical data sets available to date. We present our photometric and spectroscopic observations of this event starting several weeks before and out to approximately two years after the g-band peak brightness and combine them with public photometric data. This extensive data set robustly reveals a change in the light-curve slope and a possible bump in the rising light curve of a TDE for the first time, which may indicate more than one dominant emission mechanism contributing to the pre-peak light curve. Indeed, we find that the MOSFiT-derived parameters of AT 2019azh, which assume reprocessed accretion as the sole source of emission, are not entirely self-consistent. We further confirm the relation seen in previous TDEs whereby the redder emission peaks later than the bluer emission. The post-peak bolometric light curve of AT 2019azh is better described by an exponential decline than by the canonical t^{-5/3} (and in fact any) power-law decline. We find a possible mid-infrared excess around the peak optical luminosity, but cannot determine its origin. In addition, we provide the earliest measurements of the Halpha emission-line evolution and find no significant time delay between the peak of the V-band light curve and that of the Halpha luminosity. These results can be used to constrain future models of TDE line formation and emission mechanisms in general. More pre-peak 1-2 days cadence observations of TDEs are required to determine whether the characteristics observed here are common among TDEs. More importantly, detailed emission models are needed to fully exploit such observations for understanding the emission physics of TDEs., Comment: Submitted to ApJ

Beamforming is a signal processing technique where an array of antenna elements can be steered to transmit and receive radio signals in a specific direction. The usage of millimeter wave (mmWave) frequencies and multiple input multiple output (MIMO) beamforming are considered as the key innovations of 5th Generation (5G) and beyond communication systems. The technique initially performs a beam alignment procedure, followed by data transfer in the aligned directions between the transmitter and the receiver. Traditionally, beam alignment involves periodical and exhaustive beam sweeping at both transmitter and the receiver, which is a slow process causing extra communication overhead with MIMO and massive MIMO radio units. In applications such as beam tracking, angular velocity, beam steering etc., the beam alignment procedure is optimized by estimating the beam directions using first order polynomial approximations. Recent learning-based SOTA strategies for fast mmWave beam alignment also require exploration over exhaustive beam pairs during the training procedure, causing overhead to learning strategies for higher antenna configurations. In this work, we first optimize the beam alignment cost functions e.g. the data rate, to reduce the beam sweeping overhead by applying polynomial approximations of its partial derivatives which can then be solved as a system of polynomial equations using well-known tools from algebraic geometry. At this point, a question arises: 'what is a good polynomial approximation?' In this work, we attempt to obtain a 'good polynomial approximation'. Preliminary experiments indicate that our estimated polynomial approximations attain a so-called sweet-spot in terms of the solver speed and accuracy, when evaluated on test beamforming problems., Comment: Accepted for publication in the SIGSAM's ACM Communications in Computer Algebra, as an extended abstract

In this paper we study the problem of estimating the semi-generalized pose of a partially calibrated camera, i.e., the pose of a perspective camera with unknown focal length w.r.t. a generalized camera, from a hybrid set of 2D-2D and 2D-3D point correspondences. We study all possible camera configurations within the generalized camera system. To derive practical solvers to previously unsolved challenging configurations, we test different parameterizations as well as different solving strategies based on the state-of-the-art methods for generating efficient polynomial solvers. We evaluate the three most promising solvers, i.e., the H51f solver with five 2D-2D correspondences and one 2D-3D correspondence viewed by the same camera inside generalized camera, the H32f solver with three 2D-2D and two 2D-3D correspondences, and the H13f solver with one 2D-2D and three 2D-3D correspondences, on synthetic and real data. We show that in the presence of noise in the 3D points these solvers provide better estimates than the corresponding absolute pose solvers.

Developing an effective automatic classifier to separate genuine sources from artifacts is essential for transient follow-ups in wide-field optical surveys. The identification of transient detections from the subtraction artifacts after the image differencing process is a key step in such classifiers, known as real-bogus classification problem. We apply a self-supervised machine learning model, the deep-embedded self-organizing map (DESOM) to this "real-bogus" classification problem. DESOM combines an autoencoder and a self-organizing map to perform clustering in order to distinguish between real and bogus detections, based on their dimensionality-reduced representations. We use 32x32 normalized detection thumbnails as the input of DESOM. We demonstrate different model training approaches, and find that our best DESOM classifier shows a missed detection rate of 6.6% with a false positive rate of 1.5%. DESOM offers a more nuanced way to fine-tune the decision boundary identifying likely real detections when used in combination with other types of classifiers, for example built on neural networks or decision trees. We also discuss other potential usages of DESOM and its limitations.

We present CG-NeRF, a cascade and generalizable neural radiance fields method for view synthesis. Recent generalizing view synthesis methods can render high-quality novel views using a set of nearby input views. However, the rendering speed is still slow due to the nature of uniformly-point sampling of neural radiance fields. Existing scene-specific methods can train and render novel views efficiently but can not generalize to unseen data. Our approach addresses the problems of fast and generalizing view synthesis by proposing two novel modules: a coarse radiance fields predictor and a convolutional-based neural renderer. This architecture infers consistent scene geometry based on the implicit neural fields and renders new views efficiently using a single GPU. We first train CG-NeRF on multiple 3D scenes of the DTU dataset, and the network can produce high-quality and accurate novel views on unseen real and synthetic data using only photometric losses. Moreover, our method can leverage a denser set of reference images of a single scene to produce accurate novel views without relying on additional explicit representations and still maintains the high-speed rendering of the pre-trained model. Experimental results show that CG-NeRF outperforms state-of-the-art generalizable neural rendering methods on various synthetic and real datasets., Comment: Accepted at IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)

The Journal of Economic Literature codes classification system (JEL) published by the American Economic Association (AEA) is the de facto standard classification system for research literature in economics. The JEL classification system is used to classify articles, dissertations, books, book reviews, and working papers in EconLit, a database maintained by the AEA. Over time, it has evolved and extended to a system with over 850 subclasses. This paper reviews the history and development of the JEL classification system, describes the current version, and provides a selective overview of its uses and applications in research. The JEL codes classification system has been adopted by several publishers, and their instructions are reviewed. There are interesting avenues for future research as the JEL classification system has been surprisingly little used in existing bibliometric and scientometric research as well as in library classification systems.

Evaluation measures have a crucial impact on the direction of research. Therefore, it is of utmost importance to develop appropriate and reliable evaluation measures for new applications where conventional measures are not well suited. Video Moment Retrieval (VMR) is one such application, and the current practice is to use R@$K,\theta$ for evaluating VMR systems. However, this measure has two disadvantages. First, it is rank-insensitive: It ignores the rank positions of successfully localised moments in the top-$K$ ranked list by treating the list as a set. Second, it binarizes the Intersection over Union (IoU) of each retrieved video moment using the threshold $\theta$ and thereby ignoring fine-grained localisation quality of ranked moments. We propose an alternative measure for evaluating VMR, called Average Max IoU (AxIoU), which is free from the above two problems. We show that AxIoU satisfies two important axioms for VMR evaluation, namely, \textbf{Invariance against Redundant Moments} and \textbf{Monotonicity with respect to the Best Moment}, and also that R@$K,\theta$ satisfies the first axiom only. We also empirically examine how AxIoU agrees with R@$K,\theta$, as well as its stability with respect to change in the test data and human-annotated temporal boundaries., Comment: Accepted by CVPR2022

We present LDP, a lightweight dense prediction neural architecture search (NAS) framework. Starting from a pre-defined generic backbone, LDP applies the novel Assisted Tabu Search for efficient architecture exploration. LDP is fast and suitable for various dense estimation problems, unlike previous NAS methods that are either computational demanding or deployed only for a single subtask. The performance of LPD is evaluated on monocular depth estimation, semantic segmentation, and image super-resolution tasks on diverse datasets, including NYU-Depth-v2, KITTI, Cityscapes, COCO-stuff, DIV2K, Set5, Set14, BSD100, Urban100. Experiments show that the proposed framework yields consistent improvements on all tested dense prediction tasks, while being $5\%-315\%$ more compact in terms of the number of model parameters than prior arts., Comment: 15 pages, 11 figures, 8 tables. arXiv admin note: substantial text overlap with arXiv:2108.11105

Capturing and faithfully rendering photo-realistic humans from novel views is a fundamental problem for AR/VR applications. While prior work has shown impressive performance capture results in laboratory settings, it is non-trivial to achieve casual free-viewpoint human capture and rendering for unseen identities with high fidelity, especially for facial expressions, hands, and clothes. To tackle these challenges we introduce a novel view synthesis framework that generates realistic renders from unseen views of any human captured from a single-view and sparse RGB-D sensor, similar to a low-cost depth camera, and without actor-specific models. We propose an architecture to create dense feature maps in novel views obtained by sphere-based neural rendering, and create complete renders using a global context inpainting model. Additionally, an enhancer network leverages the overall fidelity, even in occluded areas from the original view, producing crisp renders with fine details. We show that our method generates high-quality novel views of synthetic and real human actors given a single-stream, sparse RGB-D input. It generalizes to unseen identities, and new poses and faithfully reconstructs facial expressions. Our approach outperforms prior view synthesis methods and is robust to different levels of depth sparsity., Comment: Accepted at ECCV 2022, Project page: https://www.phongnhhn.info/HVS_Net/index.html

The Gravitational-wave Optical Transient Observer (GOTO) is an array of wide-field optical telescopes, designed to exploit new discoveries from the next generation of gravitational wave detectors (LIGO, Virgo, KAGRA), study rapidly evolving transients, and exploit multi-messenger opportunities arising from neutrino and very high energy gamma-ray triggers. In addition to a rapid response mode, the array will also perform a sensitive, all-sky transient survey with few day cadence. The facility features a novel, modular design with multiple 40-cm wide-field reflectors on a single mount. In June 2017 the GOTO collaboration deployed the initial project prototype, with 4 telescope units, at the Roque de los Muchachos Observatory (ORM), La Palma, Canary Islands. Here we describe the deployment, commissioning, and performance of the prototype hardware, and discuss the impact of these findings on the final GOTO design. We also offer an initial assessment of the science prospects for the full GOTO facility that employs 32 telescope units across two sites., Comment: 19 pages, 16 Figures, accepted for publication in Monthly Notices of the Royal Astronomical Society

The typical detection rate of $\sim1$ gamma-ray burst (GRB) per day by the \emph{Fermi} Gamma-ray Burst Monitor (GBM) provides a valuable opportunity to further our understanding of GRB physics. However, the large uncertainty of the \emph{Fermi} localization typically prevents rapid identification of multi-wavelength counterparts. We report the follow-up of 93 \emph{Fermi} GRBs with the Gravitational-wave Optical Transient Observer (GOTO) prototype on La Palma. We selected 53 events (based on favourable observing conditions) for detailed analysis, and to demonstrate our strategy of searching for optical counterparts. We apply a filtering process consisting of both automated and manual steps to 60\,085 candidates initially, rejecting all but 29, arising from 15 events. With $\approx3$ GRB afterglows expected to be detectable with GOTO from our sample, most of the candidates are unlikely to be related to the GRBs. Since we did not have multiple observations for those candidates, we cannot confidently confirm the association between the transients and the GRBs. Our results show that GOTO can effectively search for GRB optical counterparts thanks to its large field of view of $\approx40$ square degrees and its depth of $\approx20$ mag. We also detail several methods to improve our overall performance for future follow-up programs of \emph{Fermi} GRBs.

Deep neural networks have recently thrived on single image depth estimation. That being said, current developments on this topic highlight an apparent compromise between accuracy and network size. This work proposes an accurate and lightweight framework for monocular depth estimation based on a self-attention mechanism stemming from salient point detection. Specifically, we utilize a sparse set of keypoints to train a FuSaNet model that consists of two major components: Fusion-Net and Saliency-Net. In addition, we introduce a normalized Hessian loss term invariant to scaling and shear along the depth direction, which is shown to substantially improve the accuracy. The proposed method achieves state-of-the-art results on NYU-Depth-v2 and KITTI while using 3.1-38.4 times smaller model in terms of the number of parameters than baseline approaches. Experiments on the SUN-RGBD further demonstrate the generalizability of the proposed method., Comment: 11 pages, 7 figures

This paper presents a novel neural architecture search method, called LiDNAS, for generating lightweight monocular depth estimation models. Unlike previous neural architecture search (NAS) approaches, where finding optimized networks are computationally highly demanding, the introduced novel Assisted Tabu Search leads to efficient architecture exploration. Moreover, we construct the search space on a pre-defined backbone network to balance layer diversity and search space size. The LiDNAS method outperforms the state-of-the-art NAS approach, proposed for disparity and depth estimation, in terms of search efficiency and output model performance. The LiDNAS optimized models achieve results superior to compact depth estimation state-of-the-art on NYU-Depth-v2, KITTI, and ScanNet, while being 7%-500% more compact in size, i.e the number of model parameters., Comment: 11 pages, 10 figures

In this paper, we propose the first minimal solutions for estimating the semi-generalized homography given a perspective and a generalized camera. The proposed solvers use five 2D-2D image point correspondences induced by a scene plane. One of them assumes the perspective camera to be fully calibrated, while the other solver estimates the unknown focal length together with the absolute pose parameters. This setup is particularly important in structure-from-motion and image-based localization pipelines, where a new camera is localized in each step with respect to a set of known cameras and 2D-3D correspondences might not be available. As a consequence of a clever parametrization and the elimination ideal method, our approach only needs to solve a univariate polynomial of degree five or three. The proposed solvers are stable and efficient as demonstrated by a number of synthetic and real-world experiments., Comment: Accepted to ICCV 2021 and to appear in the conference proceedings

In this paper, we propose enhancing monocular depth estimation by adding 3D points as depth guidance. Unlike existing depth completion methods, our approach performs well on extremely sparse and unevenly distributed point clouds, which makes it agnostic to the source of the 3D points. We achieve this by introducing a novel multi-scale 3D point fusion network that is both lightweight and efficient. We demonstrate its versatility on two different depth estimation problems where the 3D points have been acquired with conventional structure-from-motion and LiDAR. In both cases, our network performs on par with state-of-the-art depth completion methods and achieves significantly higher accuracy when only a small number of points is used while being more compact in terms of the number of parameters. We show that our method outperforms some contemporary deep learning based multi-view stereo and structure-from-motion methods both in accuracy and in compactness., Comment: 10 pages, 9 figures

We propose a new cascaded architecture for novel view synthesis, called RGBD-Net, which consists of two core components: a hierarchical depth regression network and a depth-aware generator network. The former one predicts depth maps of the target views by using adaptive depth scaling, while the latter one leverages the predicted depths and renders spatially and temporally consistent target images. In the experimental evaluation on standard datasets, RGBD-Net not only outperforms the state-of-the-art by a clear margin, but it also generalizes well to new scenes without per-scene optimization. Moreover, we show that RGBD-Net can be optionally trained without depth supervision while still retaining high-quality rendering. Thanks to the depth regression network, RGBD-Net can be also used for creating dense 3D point clouds that are more accurate than those produced by some state-of-the-art multi-view stereo methods., Comment: 19 pages, 15 figures. Code will be available at: https://github.com/phongnhhn92/RGBDNet

We present anisotropy results for anomalous cosmic-ray (ACR) protons in the energy range $\sim$0.5-35 MeV from Cosmic Ray Subsytem (CRS) data collected during calibration roll maneuvers for the magnetometer instrument when Voyager 2 (V2) was in the inner heliosheath. We use a new technique to derive for the first time the radial component of the anisotropy vector from CRS data. We find that the CRS-derived radial solar wind speeds, when converted from the radial components of the anisotropy vectors via the Compton-Getting (C-G) effect, generally agree with those similarly-derived speeds from the Low-Energy Charged Particle experiment using 28-43 keV data. However, they often differ significantly from the radial solar wind speeds measured directly by the Plasma Science (PLS) instrument. There are both periods when the C-G-derived radial solar wind speeds are significantly higher than those measured by PLS and times when they are significantly lower. The differences are not expected nor explained, but it appears that after a few years in the heliosheath the V2 radial solar wind speeds derived from the C-G method underestimate the true speeds as the spacecraft approaches the heliopause. We discuss the implications of this observation for the stagnation region reported along the Voyager 1 trajectory as it approached the heliopause inferred using the C-G method.

We describe a three-dimensional crystalline topological insulator (TI) phase of matter that exhibits spontaneous polarization. This polarization results from the presence of (approximately) flat bands on the surface of such TIs. These flat bands are a consequence of the bulk-boundary correspondence of polarized topological media, and contrary to related nodal line semimetal phases also containing surface flat bands, they span the entire surface Brillouin zone. We also present an example Hamiltonian exhibiting a Lifshitz transition from the nodal line phase to the TI phase with polarization. Utilizing elasticity tetrads, we show a complete classification of 3D crystalline TI phases and invariants. The phase with polarization naturally arises from this classification as a dual to the previously better-known 3D TI phase exhibiting quantum (spin) Hall effect. Besides polarization, another implication of the large surface flat band is the susceptibility to interaction effects such as superconductivity: the mean-field critical temperature is proportional to the size of the flat bands, and this type of systems may hence exhibit superconductivity with a very high critical temperature., Comment: 10 pages, 4 figures; new presentation and discussion

We present the discovery with Keck of the extremely infrared (IR) luminous transient AT 2017gbl, coincident with the Northern nucleus of the luminous infrared galaxy (LIRG) IRAS 23436+5257. Our extensive multi-wavelength follow-up spans ~900 days, including photometry and spectroscopy in the optical and IR, and (very long baseline interferometry) radio and X-ray observations. Radiative transfer modelling of the host galaxy spectral energy distribution and long-term pre-outburst variability in the mid-IR indicate the presence of a hitherto undetected dust obscured active galactic nucleus (AGN). The optical and near-IR spectra show broad 2000 km/s hydrogen, He I and O I emission features that decrease in flux over time. Radio imaging shows a fast evolving compact source of synchrotron emission spatially coincident with AT 2017gbl. We infer a lower limit for the radiated energy of 7.3 x 10^50 erg from the IR photometry. An extremely energetic supernova would satisfy this budget, but is ruled out by the radio counterpart evolution. Instead, we propose AT 2017gbl is related to an accretion event by the central supermassive black hole, where the spectral signatures originate in the AGN broad line region and the IR photometry is consistent with re-radiation by polar dust. Given the fast evolution of AT 2017gbl, we deem a tidal disruption event (TDE) of a star a more plausible scenario than a dramatic change in the AGN accretion rate. This makes AT 2017gbl the third TDE candidate to be hosted by a LIRG, in contrast to the so far considered TDE population discovered at optical wavelengths and hosted preferably by post-starburst galaxies., Comment: Accepted for publication by MNRAS. 32 pages, 22 figures

This paper addresses the problem of novel view synthesis by means of neural rendering, where we are interested in predicting the novel view at an arbitrary camera pose based on a given set of input images from other viewpoints. Using the known query pose and input poses, we create an ordered set of observations that leads to the target view. Thus, the problem of single novel view synthesis is reformulated as a sequential view prediction task. In this paper, the proposed Transformer-based Generative Query Network (T-GQN) extends the neural-rendering methods by adding two new concepts. First, we use multi-view attention learning between context images to obtain multiple implicit scene representations. Second, we introduce a sequential rendering decoder to predict an image sequence, including the target view, based on the learned representations. Finally, we evaluate our model on various challenging datasets and demonstrate that our model not only gives consistent predictions but also doesn't require any retraining for finetuning., Comment: Code is available at: https://github.com/phongnhhn92/TransformerGQN; Supplementary material: https://bit.ly/3kEgnzU

Recovering the scene depth from a single image is an ill-posed problem that requires additional priors, often referred to as monocular depth cues, to disambiguate different 3D interpretations. In recent works, those priors have been learned in an end-to-end manner from large datasets by using deep neural networks. In this paper, we propose guiding depth estimation to favor planar structures that are ubiquitous especially in indoor environments. This is achieved by incorporating a non-local coplanarity constraint to the network with a novel attention mechanism called depth-attention volume (DAV). Experiments on two popular indoor datasets, namely NYU-Depth-v2 and ScanNet, show that our method achieves state-of-the-art depth estimation results while using only a fraction of the number of parameters needed by the competing methods., Comment: 30 pages

Since crossing the heliopause on August 25, 2012, Voyager 1 observed reductions in galactic cosmic ray count rates caused by a time-varying depletion of particles with pitch angles near 90-deg, while intensities of particles with other pitch angles remain unchanged. Between late 2012 and mid-2017, three large-scale events occurred, lasting from ~100 to ~630 days. Omnidirectional and directional high-energy data from Voyager 1's Cosmic Ray Subsystem are used to report cosmic ray intensity variations. Omnidirectional (greater than ~20 MeV) proton-dominated measurements show up to a 3.8% intensity reduction. Bi-directional (greater than ~70 MeV) proton-dominated measurements taken from various spacecraft orientations provide insight about the depletion region's spatial properties. We characterize the anisotropy as a "notch" in an otherwise uniform pitch-angle distribution of varying depth and width centered about 90 degrees in pitch angle space. The notch averages 22-deg wide and 15% deep - signifying a depletion region that is broad and shallow. There are indications that the anisotropy is formed by a combination of magnetic trapping and cooling downstream of solar-induced transient disturbances in a region that is also likely influenced by the highly compressed fields near the heliopause.

Recent experimental discoveries of superconductivity and other exotic electronic states in twisted bilayer graphene (TBG) call for a reconsideration of our traditional theories of these states, usually based on the assumption of the presence of a Fermi surface. Here we show how such developments may even help us finding mechanisms of increasing the critical temperature of superconductivity towards the room temperature., Comment: Feature article submitted to Europhysics News. Added another paper in Ref. 2

The problem of predicting a novel view of the scene using an arbitrary number of observations is a challenging problem for computers as well as for humans. This paper introduces the Generative Adversarial Query Network (GAQN), a general learning framework for novel view synthesis that combines Generative Query Network (GQN) and Generative Adversarial Networks (GANs). The conventional GQN encodes input views into a latent representation that is used to generate a new view through a recurrent variational decoder. The proposed GAQN builds on this work by adding two novel aspects: First, we extend the current GQN architecture with an adversarial loss function for improving the visual quality and convergence speed. Second, we introduce a feature-matching loss function for stabilizing the training procedure. The experiments demonstrate that GAQN is able to produce high-quality results and faster convergence compared to the conventional approach., Comment: 12 pages, 4 figures, accepted for presentation at the Scandinavian Conference on Image Analysis 2019

In this paper we report a study of the isotopic composition of Li, Be, B and N, Ne nuclei from a 5 year time period beyond the heliopause using the CRS instruments on Voyager. By comparing the isotopic ratios, 15N/14N and 22Ne/20Ne outside the heliosphere as measured at Voyager, and which are found to be significantly lower than those measured at the same energy inside the heliosphere, we have provided strong evidence that cosmic rays of this energy have lost as much as 200 MeV/nuc or more in the solar modulation process. This is in accordance with the so called force field description of this overall modulation by Gleeson and Axford. The measurements at Voyager confirm that the unusual 14N and 22Ne cosmic ray source abundances relative to solar abundances made earlier inside the heliosphere extend to the lower energies not accessible from near Earth measurements. The low energy Li, Be and B nuclei, which are believed to be purely secondary nuclei, are found to have a (previously unobservable) peak in the differential intensity spectrum at ~100 MeV/nuc. This is in agreement with propagation predictions. The intensities of these nuclei are ~10-20% higher than those predicted in a propagation model with a matter path length lambda = 9 g/cm2 at these low energies. The isotopic composition of Li, Be and B nuclei is also consistent with that expected from propagation through interstellar matter., Comment: 16 PAGES, 2 TABLES, 5 FIGURES

We believe that the extreme solar modulation of 3-60 MeV Galactic electrons measured by Voyager in the heliosheath and the interpretation of this new data in terms of the rigidity dependence of the diffusion coefficient at low rigidities presented in this paper represents a major step in understanding diffusion theory as it applies to energetic particles. This description uses electron spectra measured at 5 different epochs and distances within the heliosheath. The diffusion dependence needed to explain the remarkable solar modulation effects observed for both electrons and higher rigidity protons as one progresses from the heliopause inward by ~25 AU to the termination shock really has two distinct rigidity regimes. Above a rigidity ~Pc the diffusion coefficient has a dependence ~beta P, the modulation is ~P and its magnitude increases linearly with radius in AU according to the integral of dr/K. This integral defines a potential, beta, called the modulation potential, thus explaining the proton variations. At rigidities

Studies on Voyager 1 using the CRS instrument have shown the presence of sub-MeV electrons in the interstellar medium beyond the heliopause. We believe that these electrons are the very low energy tail of the distribution of galactic GeV cosmic ray electrons produced in the galaxy. If so this observation places constraints on the origin and possible source distribution of these electrons in the galaxy. The intensities of these electrons as well as MeV protons and other higher energy electrons and nuclei have been followed outward from the Earth to beyond the heliopause during the 40 years of the Voyager mission. Among the other new features found in this study of the radial dependence of the electron intensity in the heliosphere are: 1. The heliosheath is a source of sub-MeV electrons as well as the already known anomalous cosmic rays of MeV and above, none of which appear to escape from the heliosphere because of an almost impenetrable heliopause at these lower energies; 2. Solar modulation effects are observed for these MeV electrons throughout the heliosphere. These modulation effects are particularly strong for electrons in the heliosheath and comprise over 90 percent of the observed intensity change of these electrons of 10-60 MeV between the Earth and the heliopause. Even for nuclei of 1 GV in rigidity, over 30 percent of the total intensity difference between the Earth and the LIM occurs in the heliosheath; 3. The 2 MeV protons studied here for the first time beyond the heliopause are also part of the low energy tail of the spectrum of galactic cosmic ray protons, similar to the tail noted above for sub MeV galactic cosmic ray electrons., Comment: 17 pages, 5 figures

This paper presents modeling of a compliant 2D manipulator, a so called soft hydraulic/fluidic elastomer actuator. Our focus is on fiber-Reinforced Fluidic Elastomer Actuators (RFEA) driven by a constant pressure hydraulic supply and modulated on/off valves. We present a model that not only provides the dynamics behavior of the system but also the kinematics of the actuator. In addition to that, the relation between the applied hydraulic pressure and the bending angle of the soft actuator and thus, its tip position is formulated in a systematic way. We also present a steady state model that calculates the bending angle given the fluid pressure which can be beneficial to find out the initial values of the parameters during the system identification process. Our experimental results verify and validate the performance of the proposed modeling approach both in transition and steady states. Due to its inherent simplicity, this model shall also be used in real-time control of the soft actuators., Comment: 8 pages, 10 figures

This paper presents design, fabrication and control of a compliant 2D manipulator, a so called soft actuator. Our focus is on fiber-reinforced elastomer actuators driven by a constant pressure hydraulic supply and modulated on/off valves. For a given diameters, we study the effect of four different elastomer materials and that of number of reinforcement fiber turns on forces generated by the actuator and maximum bending angles. For the rest of the study, we use polydimethylosiloxane (PDMS) with 240 fiber turns per 170mm length of actuator which withstand highest pressures and forces in our experiments. For the rest of the paper, we introduce two control methodologies. Firstly, we show that is possible to reasonably accurately control the pressure inside tube without measuring the pressure incorporating a simple linear tube model. This can be used, for example, in an inner-outer loop configuration with a PI position control to achieve high performance without the need for pressure measurement. Secondly, we experimentally show that a switching position control exhibits very good steady state accuracy and acceptable transient. Actuator tip position is measured using an external vision system. Our experiments included performance analysis of our soft manipulator while freely moving as well as when carrying a load., Comment: 7 pages, 13 Figures

We have measured the intensity and spectra of the cosmic ray secondary isotopes 2H and 3He and the primary isotopes H and 4He between 20-85 MeV/nuc during a 5 year time period after Voyager 1 (V1) crossed the heliopause. The data reported here is from the B-end high energy telescope. The ratios of the intensities of the secondary to primary spectra of these nuclei at low energies are sensitive indicators for determining the amount of matter traversed at these energies by their galactic cosmic ray progenitor nuclei, after acceleration, in this case mostly 4He nuclei. The measurements of secondary 3He abundances indicate that cosmic ray 4He of energies between 30-100 MeV/nuc have traversed between 7-9 g/cm2 of interstellar matter (90% H, 10% He) in a Leaky Box propagation model. This path length is also consistent with the production of secondary 2H nuclei between 20-50 MeV/nuc, which is also produced mainly by 4He in a LBM. The Boron abundance, studied in separate papers, is also consistent with this path length at energies >30 MeV/nuc. These secondary intensities imply that the interstellar cosmic ray path length may be described in a LBM in a manner consistent with a mean path length,gamma= 20.6 beta P-0.45 at rigidities above ~0.5 GV (30 MeV/nuc for A/Z=2 nuclei). Both 2H and Boron have an excess intensity vs. the predictions for path lengths of 9 g/cm2 below 30 MeV/nuc., Comment: 22 pages, 8 figures, 2 tables

We have obtained the energy spectra of cosmic ray He, B, C, O, Mg, S and Fe nuclei in the range 0.5-1.5 GeV/nuc and above using the penetrating particle mode of the High Energy Telescope, part of the Cosmic Ray Science (CRS) experiment on Voyagers 1 and 2. The data analysis procedures are the same as those used to obtain similar spectra from the identical V2 HET telescope while it was in the heliosphere between about 23 and 54 AU. The time period of analysis includes 4 years of data beyond the heliopause (HP). These new interstellar spectra are compared with various earlier experiments at the same energies at the Earth to determine the solar modulation parameter, phi. These new spectra are also compared with recent measurements of the spectra of the same nuclei measured by the same telescope at low energies. It is found that the ratio of intensities at 100 MeV/nuc to those at 1.0 GeV/nuc are significantly Z dependent. Some of this Z dependence can be explained by the Z2 dependence of energy loss by ionization in the 7-10 g/cm2 of interstellar H and He traversed by cosmic rays of these energies in the galaxy; some by the Z dependent loss due to nuclear interactions in this same material; some by possible differences in the source spectra of these nuclei and some by the non-uniformity of the source distribution and propagation conditions. The observed features of the spectra, also including a Z dependence of the peak intensities of the various nuclei, pose interesting problems related to the propagation and source distribution of these cosmic rays., Comment: 35 pages, 5 tables, 14 figures

We have derived background corrected intensities of 3-50 MeV galactic electrons observed by Voyager 1 as it passes through the heliosheath from 95 to 122 AU. The overall intensity change of the background corrected data from the inner to the outer boundary of the heliosheath is a maximum of a factor ~100 at 15 MeV. At lower energies this fractional change becomes less and the corrected electron spectra in the heliosheath becomes progressively steeper, reaching values ~ -2.5 for the spectral index just outside of the termination shock. At higher energies the spectra of electrons has an exponent changing from the negative LIS spectral index of -1.3 to values approaching zero in the heliosheath as a result of the solar modulation of the galactic electron component. The large modulation effects observed below ~100 MV are possible evidence for enhanced diffusion as part of the modulation process for electrons in the heliosheath., Comment: 28 pages, 2 tables, 13 figures

We review the recent progress in understanding the properties of spin-split superconductors under non-equilibrium conditions. Recent experiments and theories demonstrate a rich variety of transport phenomena occurring in devices based on such materials that suggest direct applications in thermoelectricity, low-dissipative spintronics, radiation detection and sensing. We discuss different experimental situations and present a theoretical framework based on quantum kinetic equations. Within this framework we provide an accurate description of the non-equilibrium distribution of charge, spin and energy, which are the relevant non-equilibrium modes, in different hybrid structures. We also review experiments on spin-split superconductors and show how transport measurements reveal the properties of the non-equilibrium modes and their mutual coupling. We discuss in detail spin injection and diffusion and very large thermoelectric effects in spin-split superconductors., Comment: Version published in Rev. Mod. Phys. Long version to be published separately elsewhere

We have determined the solar modulation potential, phi, vs. time that is observed at Voyager 1 and 2 from measurements of the H and He nuclei intensities at a common energy of 250 MeVnuc. The H nuclei have a rigidity 0.7 GV, the He nuclei 1.4 GV. These measurements cover a 40 year time period, which includes almost 4 cycles of solar 11 year sunspot variations, throughout the inner heliosphere out to the HTS at distances of 95 AU and 85 AU, respectively at V1 and V2, and then beyond in the heliosheath. Inside the HTS the modulation potential vs. time curves at V1 and V2 show a very similar temporal structure to those observed at the Earth. During a later period of maximum solar modulation from 2000.0 to 2005.0 when V1 and V2 are in the outer heliosphere between 60-94 AU, the main temporal features of the modulation potential curves at all 3 locations match up with appropriate time delays at V1 and V2 if it is assumed that spatially coherent structures are moving outward past V1 and V2, with outward speeds of up to 700 Kms negative 1. After 2004.0 V1 and V2 are at latitudes of positive 35 and negative 30 respectively, placing lower limits on the latitude extent of these structures. Beyond the HTS in the heliosheath the modulation potential slowly decreases at both spacecraft with only a weak evidence of the unusual modulation minimum observed at the Earth in 2009, for example. A sudden decrease of the modulation potential 50 MV for both H and He nuclei occurs at V1 just before the heliopause crossing at about 122 AU. This decrease has not yet been observed at V2, which is now at 113 AU and still observing a modulation potential 60 MV., Comment: 28 pages, 9 Figures

Three-dimensional topological semimetals come in different variants, either containing Weyl points or Dirac lines. Here we describe a more complicated momentum-space topological defect where several separate Dirac lines connect with each other, forming a momentum-space equivalent of the real-space nexus considered before for Helium-3. Close to the nexus the Dirac lines exhibit a transition from type I to type II lines. We consider a general model of stacked honeycomb lattices with the symmetry of Bernal (AB) stacked graphite and show that the structural mirror symmetries in such systems protect the presence of the Dirac lines, and also naturally lead to the formation of the nexus. By the bulk-boundary correspondence of topological media, the presence of Dirac lines lead to the formation of drumhead surface states at the side surfaces of the system. We calculate the surface state spectrum, and especially illustrate the effect of the nexus on these states.

We study the effect of the intrinsic (Rashba or Dresselhaus) spin-orbit interaction in superconductor--nanowire--superconductor (SNS) weak links in the presence of a spin-splitting field that can result either from an intrinsic exchange field or the Zeeman effect of an applied field. We solve the full non-linear Usadel equations numerically and analyze the resulting supercurrent through the weak link and the behavior of the density of states in the center of the wire. We point out how the presence of the spin-orbit interaction gives rise to a long-range spin triplet supercurrent, which remains finite even in the limit of very large exchange fields. In particular, we show how rotating the field leads to a sequence of transitions between the 0 and $\pi$ states as a function of the angle between the exchange field and the spin-orbit field. Simultaneously, the triplet pairing leads to a zero-energy peak in the density of states. We proceed by solving the linearized Usadel equations, showing the correspondence to the solutions of the full equations and detail the emergence of the long-range supercurrent components. Our studies are relevant for on-going investigations of supercurrent in semiconductor nanowires in the limit of several channels and in the presence of disorder., Comment: 11 pages, 10 figures; updated discussion of the angular dependence of the supercurrent

We consider the $Z_2$ topology of the Dirac lines, i.e., lines of band contacts, on an example of graphite. Four lines --- three with topological charge $N_1=1$ each and one with $N_1=-1$ --- merge together near the H-point and annihilate due to summation law $1+1+1-1=0$. The merging point is similar to the real-space nexus, an analog of the Dirac monopole at which the $Z_2$ strings terminate., Comment: 7 pages, 3 figures; some minor corrections, added references

Superconductivity is traditionally viewed as a low-temperature phenomenon. Within the BCS theory this is understood to result from the fact that the pairing of electrons takes place only close to the usually two-dimensional Fermi surface residing at a finite chemical potential. Because of this, the critical temperature is exponentially suppressed compared to the microscopic energy scales. On the other hand, pairing electrons around a dispersionless (flat) energy band leads to very strong superconductivity, with a mean-field critical temperature linearly proportional to the microscopic coupling constant. The prize to be paid is that flat bands can generally be generated only on surfaces and interfaces, where high-temperature superconductivity would show up. The flat-band character and the low dimensionality also mean that despite the high critical temperature such a superconducting state would be subject to strong fluctuations. Here we discuss the topological and non-topological flat bands discussed in different systems, and show that graphite is a good candidate for showing high-temperature flat-band interface superconductivity., Comment: Submitted as a chapter to the book on "Basic Physics of functionalized Graphite", 21 pages, 12 figures

Coupling electromagnetic waves in a cavity and mechanical vibrations via the radiation pressure of the photons [1,2] is a promising platform for investigations of quantum mechanical properties of motion of macroscopic bodies and thereby the limits of quantum mechanics [3,4]. A drawback is that the effect of one photon tends to be tiny, and hence one of the pressing challenges is to substantially increase the interaction strength towards the scale of the cavity damping rate. A novel scenario is to introduce into the setup a quantum two-level system (qubit), which, besides strengthening the coupling, allows for rich physics via strongly enhanced nonlinearities [5-8]. Addressing these issues, here we present a design of cavity optomechanics in the microwave frequency regime involving a Josephson junction qubit. We demonstrate boosting of the radiation pressure interaction energy by six orders of magnitude, allowing to approach the strong coupling regime, where a single quantum of vibrations shifts the cavity frequency by more than its linewidth. We observe nonlinear phenomena at single-photon energies, such as an enhanced damping due to the two-level system. This work opens up nonlinear cavity optomechanics as a plausible tool for the study of quantum properties of motion., Comment: 5 pages + supplementary

We show that the breakdown of the Wiedemann-Franz law due to electron--electron scattering in diffusive spin valves may result in a strong suppression of the Fano factor that describes the ratio between shot noise and average current. In the parallel configuration of magnetizations, we find the universal value $\sqrt{3}/4$ in the absence of a normal-metal spacer layer, but including the spacer leads to a non-monotonous suppression of this value before reaching back to the universal value for large spacer lengths. On the other hand, in the case of an antiparallel configuration with a negligibly small spacer, the Fano factor is $\sqrt{3 (1-P^2)}/4$, where $P$ denotes the polarization of the conductivities. For $P\rightarrow \pm 1$, the current through the system is almost noiseless., Comment: 5+ pages, 5 figures, including a short appendix

At the Voyager 1 spacecraft in the outer heliosphere, after a series of complex intensity changes starting at about May 8th, the intensities of both anomalous cosmic rays (ACR) and galactic cosmic rays (GCR) changed suddenly and decisively on August 25th (121.7 AU from the Sun). The ACR started the intensity decrease with an initial e-folding rate of intensity decrease of ~1 day. Within a matter of a few days, the intensity of 1.9-2.7 MeV protons and helium nuclei had decreased to less than 0.1 of their previous value and after a few weeks, corresponding to the outward movement of V1 by ~0.1 AU, these intensities had decreased by factors of at least 300-500 and are now lower than most estimates of the GCR spectrum for these lower energies and also at higher energies. The decrease was accompanied by large rigidity dependent anisotropies in addition to the extraordinary rapidity of the intensity changes. Also on August 25th the GCR protons, helium and heavier nuclei as well as electrons increased suddenly with the intensities of electrons reaching levels ~30-50% higher than observed just one day earlier. This increase for GCR occurred over ~1 day for the lowest rigidity electrons, and several days for the higher rigidity nuclei of rigidity ~0.5-1.0 GV. After reaching these higher levels the intensities of the GCR of all energies from 2 to 400 MeV have remained essentially constant with intensity levels and spectra that may represent the local GCR. These intensity changes will be presented in more detail in this, and future articles, as this story unfolds., Comment: 13 Pages, 5 Figures