
Multilevel partitioning methods that are inspired by principles of
multiscaling are the most powerful practical hypergraph partitioning solvers.
Hypergraph partitioning has many applications in disciplines ranging from
scientific computing to data science. In this paper we introduce the concept of
algebraic distance on hypergraphs and demonstrate its use as an algorithmic
component in the coarsening stage of multilevel hypergraph partitioning
solvers. The algebraic distance is a vertex distance measure that extends
hyperedge weights for capturing the local connectivity of vertices which is
critical for hypergraph coarsening schemes. The practical effectiveness of the
proposed measure and corresponding coarsening scheme is demonstrated through
extensive computational experiments on a diverse set of problems. Finally, we
propose a benchmark of hypergraph partitioning problems to compare the quality
of other solvers.

In today's cyberenabled smart grids, high penetration of uncertain
renewables, purposeful manipulation of meter readings, and the need for
widearea situational awareness, call for fast, accurate, and robust power
system state estimation. The leastabsolutevalue (LAV) estimator is known for
its robustness relative to the weighted leastsquares (WLS) one. However, due
to nonconvexity and nonsmoothness, existing LAV solvers based on linear
programming are typically slow, hence inadequate for realtime system
monitoring. This paper develops two novel algorithms for efficient LAV
estimation, which draw from recent advances in composite optimization. The
first is a deterministic linear proximal scheme that handles a sequence of
convex quadratic problems, each efficiently solvable either via offtheshelf
algorithms or through the alternating direction method of multipliers.
Leveraging the sparse connectivity inherent to power networks, the second
scheme is stochastic, and updates only \emph{a few} entries of the complex
voltage state vector per iteration. In particular, when voltage magnitude and
(re)active power flow measurements are used only, this number reduces to one or
two, \emph{regardless of} the number of buses in the network. This
computational complexity evidently scales well to largesize power systems.
Furthermore, by carefully \emph{minibatching} the voltage and power flow
measurements, accelerated implementation of the stochastic iterations becomes
possible. The developed algorithms are numerically evaluated using a variety of
benchmark power networks. Simulated tests corroborate that improved robustness
can be attained at comparable or markedly reduced computation times for medium
or largesize networks relative to the "workhorse" WLSbased GaussNewton
iterations.

Using Blanchfield pairings, we show that two Alexander polynomials cannot be
realized by a pair of matrices with Gordian distance one if a corresponding
quadratic equation does not have an integer solution. We also give an example
of how our results help in calculating the Gordian distances, algebraic Gordian
distances and polynomial distances.

We consider the estimation of the state transition matrix in vector
autoregressive models, when time sequence data is limited but nonsequence
steadystate data is abundant. To leverage both sources of data, we formulate
the least squares minimization problem regularized by a Lyapunov penalty. We
impose cardinality or rank constraints to reduce the complexity of the
autoregressive model. We solve the resulting nonconvex, nonsmooth problem by
using the proximal alternating linearization method (PALM). We show that PALM
is globally convergent to a critical point and that the estimation error
monotonically decreases. Furthermore, we obtain explicit formulas for the
proximal operators to facilitate the implementation of PALM. We demonstrate the
effectiveness of the developed method on synthetic and realworld data. Our
experiments show that PALM outperforms the gradient projection method in both
computational efficiency and solution quality.

Saliency integration has attracted much attention on unifying saliency maps
from multiple saliency models. Previous offline integration methods usually
face two challenges: 1. if most of the candidate saliency models misjudge the
saliency on an image, the integration result will lean heavily on those
inferior candidate models; 2. an unawareness of the ground truth saliency
labels brings difficulty in estimating the expertise of each candidate model.
To address these problems, in this paper, we propose an arbitrator model (AM)
for saliency integration. Firstly, we incorporate the consensus of multiple
saliency models and the external knowledge into a reference map to effectively
rectify the misleading by candidate models. Secondly, our quest for ways of
estimating the expertise of the saliency models without ground truth labels
gives rise to two distinct online modelexpertise estimation methods. Finally,
we derive a Bayesian integration framework to reconcile the saliency models of
varying expertise and the reference map. To extensively evaluate the proposed
AM model, we test twentyseven stateoftheart saliency models, covering both
traditional and deep learning ones, on various combinations over four datasets.
The evaluation results show that the AM model improves the performance
substantially compared to the existing stateoftheart integration methods,
regardless of the chosen candidate saliency models.

This paper investigates the distributed computation of the wellknown linear
matrix equation in the form of $AXB = F$, with the matrices A, B, X, and F of
appropriate dimensions, over multiagent networks from an optimization
perspective. In this paper, we consider the standard distributed
matrixinformation structures, where each agent of the considered multiagent
network has access to one of the subblock matrices of A, B, and F. To be
specific, we first propose different decomposition methods to reformulate the
matrix equations in standard structures as distributed constrained optimization
problems by introducing substitutional variables; we show that the solutions of
the reformulated distributed optimization problems are equivalent to least
squares solutions to original matrix equations; and we design distributed
continuoustime algorithms for the constrained optimization problems, even by
using augmented matrices and a derivative feedback technique. Moreover, we
prove the exponential convergence of the algorithms to a least squares solution
to the matrix equation for any initial condition.

In this paper, we consider the robustness of a basic model of a dynamical
distribution network. In the first problem, i.e., optimal weight allocation, we
minimize the Hinf norm of the dynamical distribution network subject to
allocation of the weights on the edges. It is shown that this optimization
problem can be formulated as a semidefinite program. Next we consider the
semidefiniteness of the weighted graph Laplacian matrix with negative weights
on the edges. A necessary and sufficient condition, using the effective
resistance matrix, is established to guarantee the positive semidefiniteness
of the Laplacian matrix. Furthermore, the bounded real lemma is derived for
statespace symmetric systems.

Outdoor visionbased systems suffer from atmospheric turbulences, and rain is
one of the worst factors for vision degradation. Current rain removal methods
show limitations either for complex dynamic scenes, or under torrential rain
with opaque occlusions. We propose a novel derain framework which applies
superpixel (SP) segmentation to decompose the scene into depth consistent
units. Alignment of scene contents are done at the SP level, which proves to be
robust against rain occlusion interferences and fast camera motion. Two
alignment output tensors, i.e., optimal temporal match tensor and sorted
spatialtemporal match tensor, provide informative clues for the location of
rain streaks and the occluded background contents. Different classical and
novel methods such as Robust Principle Component Analysis and Convolutional
Neural Networks are applied and compared for their respective advantages in
efficiently exploiting the rich spatialtemporal features provided by the two
tensors. Extensive evaluations show that advantage of up to 5dB is achieved on
the scene restoration PSNR over stateoftheart methods, and the advantage is
especially obvious with highly complex and dynamic scenes. Visual evaluations
show that the proposed framework is not only able to suppress heavy and opaque
occluding rain streaks, but also large semitransparent regional fluctuations
and distortions.

Biexcitons are a manifestation of manybody excitonic interactions crucial
for quantum information and quantum computation in the construction of coherent
combinations of quantum states. However, due to their small binding energy and
low transition efficiency, most biexcitons in conventional semiconductors exist
either at cryogenic temperature or under femtosecond pulse laser excitation.
Here we demonstrate room temperature, continuous wave driven biexciton states
in CsPbBr3 perovskite nanocrystals through coupling with a plasmonic nanogap.
The room temperature CsPbBr3 biexciton excitation fluence (~100 mW/cm2) is
reduced by ~10^13 times in the Ag nanowirefilm nanogaps. The giant enhancement
of biexciton emission is driven by coherent biexcitonplasmon Fano
interference. These results provide new pathways to develop high efficiency
nonblinking single photon sources, entangled light sources and lasers based on
biexciton states.

In layered LiNixMnyCozO2 cathode material for lithiumion batteries, the
spins of transition metal (TM) ions construct a twodimensional triangular
networks, which can be considered as a simple case of geometrical frustration.
By performing neutron powder diffraction experiments and magnetization
measurements, we find that longrange magnetic order cannot be established in
LiNixMnyCozO2 even at low temperature of 3 K. Remarkably, the frustration
parameters of these compounds are estimated to be larger than 30, indicating
the existence of strongly frustrated magnetic interactions between spins of TM
ions. As frustration will inevitably give rise to lattice instability, the
formation of Li/Ni exchange in LiNixMnyCozO2 will help to partially relieve the
degeneracy of the frustrated magnetic lattice by forming a stable
antiferromagnetic state in hexagonal sublattice with nonmagnetic ions located
in centers of the hexagons. Moreover, Li/Ni exchange will introduce 180{\deg}
superexchange interaction, which further relieves the magnetic frustration
through bringing in new exchange paths. Thus, the variation of Li/Ni exchange
ratio vs. TM mole fraction in LiNixMnyCozO2 with different compositions can be
well understood and predicted in terms of magnetic frustration and
superexchange interactions. This provides a unique viewpoint to study the Li/Ni
ions exchange in layered Li(NixMnyCoz)O2 cathode materials.

Group zeroattracting LMS and its reweighted form have been proposed for
addressing system identification problems with structural group sparsity in the
parameters to estimate. Both algorithms however suffer from a tradeoff between
sparsity degree and estimation bias and, in addition, between convergence speed
and steadystate performance like most adaptive filtering algorithms. It is
therefore necessary to properly set their step size and regularization
parameter. Based on a model of their transient behavior, we introduce a
variableparameter variant of both algorithms to address this issue. By
minimizing their meansquare deviation at each time instant, we obtain
closedform expressions of the optimal step size and regularization parameter.
Simulation results illustrate the effectiveness of the proposed algorithms.

Rain removal is important for improving the robustness of outdoor vision
based systems. Current rain removal methods show limitations either for complex
dynamic scenes shot from fast moving cameras, or under torrential rain fall
with opaque occlusions. We propose a novel derain algorithm, which applies
superpixel (SP) segmentation to decompose the scene into depth consistent
units. Alignment of scene contents are done at the SP level, which proves to be
robust towards rain occlusion and fast camera motion. Two alignment output
tensors, i.e., optimal temporal match tensor and sorted spatialtemporal match
tensor, provide informative clues for rain streak location and occluded
background contents to generate an intermediate derain output. These tensors
will be subsequently prepared as input features for a convolutional neural
network to restore high frequency details to the intermediate output for
compensation of misalignment blur. Extensive evaluations show that up to 5 dB
reconstruction PSNR advantage is achieved over stateoftheart methods. Visual
inspection shows that much cleaner rain removal is achieved especially for
highly dynamic scenes with heavy and opaque rainfall from a fast moving camera.

Ultrafast lattice deformation of tens to hundreds of nanometer thick metallic
crystals, after femtosecond laser excitation, was measured directly using 8.04
keV subpicosecond xray and 59 keV femtosecond electron pulses. Coherent
phonons were generated in both single crystal and polycrystalline films.
Lattice compression was observed within the first few picoseconds after laser
irradiation in single crystal aluminum, which was attributed to the generation
of a blast force and the propagation of elastic waves. The different time scale
of lattice heating for tens and hundreds nanometer thick films are clearly
distinguished by electron and xray pulse diffraction. The electron and lattice
heating due to ultrafast deposition of photon energy was numerically simulated
using the twotemperature model (TTM) and the results agreed with experimental
observations. The ultrafast heating described by TTM was also discussed from an
electrical circuit perspective, which may provide new insights on the possible
connection between thermal and electrical processes. This study demonstrates
that the combination of two complimentary ultrafast timeresolved methods,
ultrafast xray and electron diffraction will provide a panoramic picture of
the transient atomic motions and structure in crystals.

Modifying phonon thermal conductivity in nanomaterials is important not only
for fundamental research but also for practical applications. However, the
experiments on tailoring the thermal conductivity in nanoscale, especially in
twodimensional materials, are rare due to technical challenges. In this work,
we demonstrate insitu thermal conduction measurement of MoS2 and find that its
thermal conductivity can be continuously tuned to a required value from
crystalline to amorphous limits. The reduction of thermal conductivity is
understood from phonondefects scatterings that decrease the phonon
transmission coefficient. Beyond a threshold, a sharp drop in thermal
conductivity is observed, which is believed to be a crystallineamorphous
transition. Our method and results provide guidance for potential applications
in thermoelectrics, photoelectronics, and energy harvesting where thermal
management is critical with further integration and miniaturization.

In this paper, we design and analyze a new zerothorder online algorithm,
namely, the zerothorder online alternating direction method of multipliers
(ZOOADMM), which enjoys dual advantages of being gradientfree operation and
employing the ADMM to accommodate complex structured regularizers. Compared to
the firstorder gradientbased online algorithm, we show that ZOOADMM requires
$\sqrt{m}$ times more iterations, leading to a convergence rate of
$O(\sqrt{m}/\sqrt{T})$, where $m$ is the number of optimization variables, and
$T$ is the number of iterations. To accelerate ZOOADMM, we propose two
minibatch strategies: gradient sample averaging and observation averaging,
resulting in an improved convergence rate of $O(\sqrt{1+q^{1}m}/\sqrt{T})$,
where $q$ is the minibatch size. In addition to convergence analysis, we also
demonstrate ZOOADMM to applications in signal processing, statistics, and
machine learning.

An outstanding problem when computing a function of a matrix, $f(A)$, by
using a Krylov method is to accurately estimate errors when convergence is
slow. Apart from the case of the exponential function which has been
extensively studied in the past, there are no wellestablished solutions to the
problem. Often the quantity of interest in applications is not the matrix
$f(A)$ itself, but rather, matrixvector products or bilinear forms. When the
computation related to $f(A)$ is a building block of a larger problem (e.g.,
approximately computing its trace), a consequence of the lack of reliable error
estimates is that the accuracy of the computed result is unknown. In this
paper, we consider the problem of computing $\mathrm{tr}(f(A))$ for a symmetric
positivedefinite matrix $A$ by using the Lanczos method and make two
contributions: (i) we propose an error estimate for the bilinear form
associated with $f(A)$, and (ii) an error estimate for the trace of $f(A)$. We
demonstrate the practical usefulness of these estimates for large matrices and
in particular, show that the trace error estimate is indicative of the number
of accurate digits. As an application, we compute the logdeterminant of a
covariance matrix in Gaussian process analysis and underline the importance of
error tolerance as a stopping criterion, as a means of bounding the number of
Lanczos steps to achieve a desired accuracy.

Research in texture recognition often concentrates on recognizing textures
with intraclass variations such as illumination, rotation, viewpoint and small
scale changes. In contrast, in realworld applications a change in scale can
have a dramatic impact on texture appearance, to the point of changing
completely from one texture category to another. As a result, texture
variations due to changes in scale are amongst the hardest to handle. In this
work we conduct the first study of classifying textures with extreme variations
in scale. To address this issue, we first propose and then reduce scale
proposals on the basis of dominant texture patterns. Motivated by the
challenges posed by this problem, we propose a new GANet network where we use a
Genetic Algorithm to change the units in the hidden layers during network
training, in order to promote the learning of more informative semantic texture
patterns. Finally, we adopt a FVCNN (Fisher Vector pooling of a Convolutional
Neural Network filter bank) feature encoder for global texture representation.
Because extreme scale variations are not necessarily present in most standard
texture databases, to support the proposed extremescale aspects of texture
understanding we are developing a new dataset, the Extreme Scale Variation
Textures (ESVaT), to test the performance of our framework. It is demonstrated
that the proposed framework significantly outperforms goldstandard texture
features by more than 10% on ESVaT. We also test the performance of our
proposed approach on the KTHTIPS2b and OS datasets and a further dataset
synthetically derived from Forrest, showing superior performance compared to
the state of the art.

Texture is a fundamental characteristic of many types of images, and texture
representation is one of the essential and challenging problems in computer
vision and pattern recognition which has attracted extensive research
attention. Since 2000, texture representations based on Bag of Words (BoW) and
on Convolutional Neural Networks (CNNs) have been extensively studied with
impressive performance. Given this period of remarkable evolution, this paper
aims to present a comprehensive survey of advances in texture representation
over the last two decades. More than 200 major publications are cited in this
survey covering different aspects of the research, which includes (i) problem
description; (ii) recent advances in the broad categories of BoWbased,
CNNbased and attributebased methods; and (iii) evaluation issues,
specifically benchmark datasets and state of the art results. In retrospect of
what has been achieved so far, the survey discusses open challenges and
directions for future research.

The graph convolutional networks (GCN) recently proposed by Kipf and Welling
are an effective graph model for semisupervised learning. This model, however,
was originally designed to be learned with the presence of both training and
test data. Moreover, the recursive neighborhood expansion across layers poses
time and memory challenges for training with large, dense graphs. To relax the
requirement of simultaneous availability of test data, we interpret graph
convolutions as integral transforms of embedding functions under probability
measures. Such an interpretation allows for the use of Monte Carlo approaches
to consistently estimate the integrals, which in turn leads to a batched
training scheme as we propose in this workFastGCN. Enhanced with importance
sampling, FastGCN not only is efficient for training but also generalizes well
for inference. We show a comprehensive set of experiments to demonstrate its
effectiveness compared with GCN and related models. In particular, training is
orders of magnitude more efficient while predictions remain comparably
accurate.

On September 10, 2017, Hurricane Irma made landfall in the Florida Keys and
caused significant damage. Informed by hydrodynamic storm surge and wave
modeling and poststorm satellite imagery, a rapid damage survey was soon
conducted for 1600+ residential buildings in Big Pine Key and Marathon. Damage
categorizations and statistical analysis reveal distinct factors governing
damage at these two locations. The distance from the coast is significant for
the damage in Big Pine Key, as severely damaged buildings were located near
narrow waterways connected to the ocean. Building type and size are critical in
Marathon, highlighted by the nearcomplete destruction of trailer communities
there. These observations raise issues of affordability and equity that need
consideration in damage recovery and rebuilding for resilience.

We present results from the investigation of 5min umbral oscillations in a
singlepolarity sunspot of active region NOAA 12132. The spectra of TiO,
H$\alpha$, and 304 \AA{} are used for corresponding atmospheric heights from
the photosphere to lower corona. Power spectrum analysis at the formation
height of H$\alpha$  0.6 \AA{} to H$\alpha$ center resulted in the detection
of 5min oscillation signals in intensity interpreted as running waves outside
the umbral center, mostly with vertical magnetic field inclination $>15\deg$. A
phasespeed filter is used to extract the running wave signals with speed
$v_{ph}> 4$ km s$^{1}$, from the time series of H$\alpha$  0.4 \AA{} images,
and found twentyfour 3min umbral oscillatory events in a duration of one
hour. Interestingly, the initial emergence of the 3min umbral oscillatory
events are noticed closer to or at umbral boundaries. These 3min umbral
oscillatory events are observed for the first time as propagating from a
fraction of preceding Running Penumbral Waves (RPWs). These fractional
wavefronts rapidly separates from RPWs and move towards umbral center, wherein
they expand radially outwards suggesting the beginning of a new umbral
oscillatory event. We found that most of these umbral oscillatory events
develop further into RPWs. We speculate that the waveguides of running waves
are twisted in spiral structures and hence the wavefronts are first seen at
high latitudes of umbral boundaries and later at lower latitudes of the umbral
center.

In this paper, we study the $H_\infty$norm of linear systems over graphs,
which is used to model distribution networks. In particular, we aim to minimize
the $H_\infty$norm subject to allocation of the weights on the edges. The
optimization problem is formulated with LMI (LinearMatrixInequality)
constraints. For distribution networks with one port, i.e., SISO systems, we
show that the $H_\infty$norm coincides with the effective resistance between
the nodes in the port. Moreover, we derive an upper bound of the
$H_\infty$norm, which is in terms of the algebraic connectivity of the graph
on which the distribution network is defined.

We propose a novel class of kernels to alleviate the high computational cost
of largescale nonparametric learning with kernel methods. The proposed kernel
is defined based on a hierarchical partitioning of the underlying data domain,
where the Nystr\"om method (a globally lowrank approximation) is married with
a locally lossless approximation in a hierarchical fashion. The kernel
maintains (strict) positivedefiniteness. The corresponding kernel matrix
admits a recursively offdiagonal lowrank structure, which allows for fast
linear algebra computations. Suppressing the factor of data dimension, the
memory and arithmetic complexities for training a regression or a classifier
are reduced from $O(n^2)$ and $O(n^3)$ to $O(nr)$ and $O(nr^2)$, respectively,
where $n$ is the number of training examples and $r$ is the rank on each level
of the hierarchy. Although other randomized approximate kernels entail a
similar complexity, empirical results show that the proposed kernel achieves a
matching performance with a smaller $r$. We demonstrate comprehensive
experiments to show the effective use of the proposed kernel on data sizes up
to the order of millions.

Excitonpolaritons in semiconductor microcavities generate fascinating
effects such as longrange spatial coherence and BoseEinstein Condensation
(BEC), which are attractive for their potential use in low threshold lasers,
vortices and slowing light, etc. However, currently most of excitonpolariton
effects either occur at cryogenic temperature or rely on expensive cavity
fabrication procedures. Further exploring new semiconductor microcavities with
stronger exciton photon interaction strength is extensively needed. Herein, we
demonstrate room temperature photon exciton strong coupling in hybrid
inorganicorganic CH3NH3PbBr3 FabryP\'erot microcavities for the first time.
The vacuum Rabi splitting energy is up to ~390 meV, which is ascribed to large
oscillator strength and photon confinement in reduced dimension of optical
microcavities. With increasing pumping energy, excitonphoton coupling strength
is weakened due to carrier screening effect, leading to occurrence of photonic
lasing instead of polartion lasing. The demonstrated strong coupling between
photons and excitons in perovskite microcavities would be helpful for
development of high performance polaritonbased incoherent and coherent light
sources, nonlinear optics, and slow light applications.

Depth estimation is a fundamental problem for light field photography
applications. Numerous methods have been proposed in recent years, which either
focus on crafting cost terms for more robust matching, or on analyzing the
geometry of scene structures embedded in the epipolarplane images. Significant
improvements have been made in terms of overall depth estimation error;
however, current stateoftheart methods still show limitations in handling
intricate occluding structures and complex scenes with multiple occlusions. To
address these challenging issues, we propose a very effective depth estimation
framework which focuses on regularizing the initial label confidence map and
edge strength weights. Specifically, we first detect partially occluded
boundary regions (POBR) via superpixel based regularization. Series of
shrinkage/reinforcement operations are then applied on the label confidence map
and edge strength weights over the POBR. We show that after weight
manipulations, even a lowcomplexity weighted least squares model can produce
much better depth estimation than stateoftheart methods in terms of average
disparity error rate, occlusion boundary precisionrecall rate, and the
preservation of intricate visual features.