title
stringlengths 7
239
| abstract
stringlengths 7
2.76k
| cs
int64 0
1
| phy
int64 0
1
| math
int64 0
1
| stat
int64 0
1
| quantitative biology
int64 0
1
| quantitative finance
int64 0
1
|
|---|---|---|---|---|---|---|---|
End-to-end semantic face segmentation with conditional random fields as convolutional, recurrent and adversarial networks | Recent years have seen a sharp increase in the number of related yet distinct
advances in semantic segmentation. Here, we tackle this problem by leveraging
the respective strengths of these advances. That is, we formulate a conditional
random field over a four-connected graph as end-to-end trainable convolutional
and recurrent networks, and estimate them via an adversarial process.
Importantly, our model learns not only unary potentials but also pairwise
potentials, while aggregating multi-scale contexts and controlling higher-order
inconsistencies. We evaluate our model on two standard benchmark datasets for
semantic face segmentation, achieving state-of-the-art results on both of them.
| 1 | 0 | 0 | 0 | 0 | 0 |
Bayesian Coreset Construction via Greedy Iterative Geodesic Ascent | Coherent uncertainty quantification is a key strength of Bayesian methods.
But modern algorithms for approximate Bayesian posterior inference often
sacrifice accurate posterior uncertainty estimation in the pursuit of
scalability. This work shows that previous Bayesian coreset construction
algorithms---which build a small, weighted subset of the data that approximates
the full dataset---are no exception. We demonstrate that these algorithms scale
the coreset log-likelihood suboptimally, resulting in underestimated posterior
uncertainty. To address this shortcoming, we develop greedy iterative geodesic
ascent (GIGA), a novel algorithm for Bayesian coreset construction that scales
the coreset log-likelihood optimally. GIGA provides geometric decay in
posterior approximation error as a function of coreset size, and maintains the
fast running time of its predecessors. The paper concludes with validation of
GIGA on both synthetic and real datasets, demonstrating that it reduces
posterior approximation error by orders of magnitude compared with previous
coreset constructions.
| 0 | 0 | 0 | 1 | 0 | 0 |
Practical Bayesian optimization in the presence of outliers | Inference in the presence of outliers is an important field of research as
outliers are ubiquitous and may arise across a variety of problems and domains.
Bayesian optimization is method that heavily relies on probabilistic inference.
This allows outstanding sample efficiency because the probabilistic machinery
provides a memory of the whole optimization process. However, that virtue
becomes a disadvantage when the memory is populated with outliers, inducing
bias in the estimation. In this paper, we present an empirical evaluation of
Bayesian optimization methods in the presence of outliers. The empirical
evidence shows that Bayesian optimization with robust regression often produces
suboptimal results. We then propose a new algorithm which combines robust
regression (a Gaussian process with Student-t likelihood) with outlier
diagnostics to classify data points as outliers or inliers. By using an
scheduler for the classification of outliers, our method is more efficient and
has better convergence over the standard robust regression. Furthermore, we
show that even in controlled situations with no expected outliers, our method
is able to produce better results.
| 1 | 0 | 0 | 1 | 0 | 0 |
Is there any polynomial upper bound for the universal labeling of graphs? | A {\it universal labeling} of a graph $G$ is a labeling of the edge set in
$G$ such that in every orientation $\ell$ of $G$ for every two adjacent
vertices $v$ and $u$, the sum of incoming edges of $v$ and $u$ in the oriented
graph are different from each other. The {\it universal labeling number} of a
graph $G$ is the minimum number $k$ such that $G$ has {\it universal labeling}
from $\{1,2,\ldots, k\}$ denoted it by $\overrightarrow{\chi_{u}}(G) $. We have
$2\Delta(G)-2 \leq \overrightarrow{\chi_{u}} (G)\leq 2^{\Delta(G)}$, where
$\Delta(G)$ denotes the maximum degree of $G$. In this work, we offer a
provocative question that is:" Is there any polynomial function $f$ such that
for every graph $G$, $\overrightarrow{\chi_{u}} (G)\leq f(\Delta(G))$?".
Towards this question, we introduce some lower and upper bounds on their
parameter of interest. Also, we prove that for every tree $T$,
$\overrightarrow{\chi_{u}}(T)=\mathcal{O}(\Delta^3) $. Next, we show that for a
given 3-regular graph $G$, the universal labeling number of $G$ is 4 if and
only if $G$ belongs to Class 1. Therefore, for a given 3-regular graph $G$, it
is an $ \mathbf{NP} $-complete to determine whether the universal labeling
number of $G$ is 4. Finally, using probabilistic methods, we almost confirm a
weaker version of the problem.
| 1 | 0 | 1 | 0 | 0 | 0 |
An Empirical Study on Team Formation in Online Games | Online games provide a rich recording of interactions that can contribute to
our understanding of human behavior. One potential lesson is to understand what
motivates people to choose their teammates and how their choices leadto
performance. We examine several hypotheses about team formation using a large,
longitudinal dataset from a team-based online gaming environment. Specifically,
we test how positive familiarity, homophily, and competence determine team
formationin Battlefield 4, a popular team-based game in which players choose
one of two competing teams to play on. Our dataset covers over two months of
in-game interactions between over 380,000 players. We show that familiarity is
an important factorin team formation, while homophily is not. Competence
affects team formation in more nuanced ways: players with similarly high
competence team-up repeatedly, but large variations in competence discourage
repeated interactions.
| 1 | 0 | 0 | 0 | 0 | 0 |
Nematic superconductivity in Cu$_{x}$Bi$_{2}$Se$_{3}$: The surface Andreev bound states | We study theoretically the topological surface states (TSSs) and the possible
surface Andreev bound states (SABSs) of Cu$_{x}$Bi$_{2}$Se$_{3}$ which is known
to be a topological insulator at $x=0$. The superconductivity (SC) pairing of
this compound is assumed to have the broken spin-rotation symmetry, similar to
that of the A-phase of $^{3}$He as suggested by recent nuclear-magnetic
resonance experiments. For both spheroidal and corrugated cylindrical Fermi
surfaces with the hexagonal warping terms, we show that the bulk SC gap is
rather anisotropic; the minimum of the gap is negligibly small as comparing to
the maximum of the gap. This would make the fully-gapped pairing effectively
nodal. For a clean system, our results indicate the bulk of this compound to be
a topological superconductor with the SABSs appearing inside the bulk SC gap.
The zero-energy SABSs which are Majorana fermions, together with the TSSs not
gapped by the pairing, produce a zero-energy peak in the surface density of
states (SDOS). The SABSs are expected to be stable against short-range
nonmagnetic impurities, and the local SDOS is calculated around a nonmagnetic
impurity. The relevance of our results to experiments is discussed.
| 0 | 1 | 0 | 0 | 0 | 0 |
The structure of a minimal $n$-chart with two crossings II: Neighbourhoods of $Γ_1\cupΓ_{n-1}$ | Given a 2-crossing minimal chart $\Gamma$, a minimal chart with two
crossings, set $\alpha=\min\{~i~|~$there exists an edge of label $i$ containing
a white vertex$\}$, and $\beta=\max\{~i~|~$there exists an edge of label $i$
containing a white vertex$\}$. In this paper we study the structure of a
neighbourhood of $\Gamma_\alpha\cup\Gamma_\beta$, and propose a normal form for
2-crossing minimal $n$-charts, here $\Gamma_\alpha$ and $\Gamma_\beta$ mean the
union of all the edges of label $\alpha$ and $\beta$ respectively.
| 0 | 0 | 1 | 0 | 0 | 0 |
Generating Shared Latent Variables for Robots to Imitate Human Movements and Understand their Physical Limitations | Assistive robotics and particularly robot coaches may be very helpful for
rehabilitation healthcare. In this context, we propose a method based on
Gaussian Process Latent Variable Model (GP-LVM) to transfer knowledge between a
physiotherapist, a robot coach and a patient. Our model is able to map visual
human body features to robot data in order to facilitate the robot learning and
imitation. In addition , we propose to extend the model to adapt robots'
understanding to patient's physical limitations during the assessment of
rehabilitation exercises. Experimental evaluation demonstrates promising
results for both robot imitation and model adaptation according to the
patients' limitations.
| 1 | 0 | 0 | 0 | 0 | 0 |
Superconvergence analysis of linear FEM based on the polynomial preserving recovery and Richardson extrapolation for Helmholtz equation with high wave number | We study superconvergence property of the linear finite element method with
the polynomial preserving recovery (PPR) and Richardson extrapolation for the
two dimensional Helmholtz equation. The $H^1$-error estimate with explicit
dependence on the wave number $k$ {is} derived.
First, we prove that under the assumption $k(kh)^2\leq C_0$ ($h$ is the mesh
size) and certain mesh condition, the estimate between the finite element
solution and the linear interpolation of the exact solution is superconvergent
under the $H^1$-seminorm, although the pollution error still exists. Second, we
prove a similar result for the recovered gradient by PPR and found that the PPR
can only improve the interpolation error and has no effect on the pollution
error. Furthermore, we estimate the error between the finite element gradient
and recovered gradient and discovered that the pollution error is canceled
between these two quantities. Finally, we apply the Richardson extrapolation to
recovered gradient and demonstrate numerically that PPR combined with the
Richardson extrapolation can reduce the interpolation and pollution errors
simultaneously, and therefore, leads to an asymptotically exact {\it a
posteriori} error estimator. All theoretical findings are verified by numerical
tests.
| 0 | 0 | 1 | 0 | 0 | 0 |
Semi-blind source separation with multichannel variational autoencoder | This paper proposes a multichannel source separation technique called the
multichannel variational autoencoder (MVAE) method, which uses a conditional
VAE (CVAE) to model and estimate the power spectrograms of the sources in a
mixture. By training the CVAE using the spectrograms of training examples with
source-class labels, we can use the trained decoder distribution as a universal
generative model capable of generating spectrograms conditioned on a specified
class label. By treating the latent space variables and the class label as the
unknown parameters of this generative model, we can develop a
convergence-guaranteed semi-blind source separation algorithm that consists of
iteratively estimating the power spectrograms of the underlying sources as well
as the separation matrices. In experimental evaluations, our MVAE produced
better separation performance than a baseline method.
| 0 | 0 | 0 | 1 | 0 | 0 |
Quantitative characterization of pore structure of several biochars with 3D imaging | Pore space characteristics of biochars may vary depending on the used raw
material and processing technology. Pore structure has significant effects on
the water retention properties of biochar amended soils. In this work, several
biochars were characterized with three-dimensional imaging and image analysis.
X-ray computed microtomography was used to image biochars at resolution of 1.14
$\mu$m and the obtained images were analysed for porosity, pore-size
distribution, specific surface area and structural anisotropy. In addition,
random walk simulations were used to relate structural anisotropy to diffusive
transport. Image analysis showed that considerable part of the biochar volume
consist of pores in size range relevant to hydrological processes and storage
of plant available water. Porosity and pore-size distribution were found to
depend on the biochar type and the structural anisotopy analysis showed that
used raw material considerably affects the pore characteristics at micrometre
scale. Therefore attention should be paid to raw material selection and quality
in applications requiring optimized pore structure.
| 0 | 1 | 0 | 0 | 0 | 0 |
First-principles-based method for electron localization: Application to monolayer hexagonal boron nitride | We present a first-principles-based many-body typical medium dynamical
cluster approximation method for characterizing electron localization in
disordered structures. This method applied to monolayer hexagonal boron nitride
shows that the presence of a boron vacancies could turn this wide-gap insulator
into a correlated metal. Depending on the strength of the electron
interactions, these calculations suggest that conduction could be obtained at a
boron vacancy concentration as low as $1.0\%$. We also explore the distribution
of the local density of states, a fingerprint of spatial variations, which
allows localized and delocalized states to be distinguished. The presented
method enables the study of disorder-driven insulator-metal transitions not
only in $h$-BN but also in other physical materials.
| 0 | 1 | 0 | 0 | 0 | 0 |
Compressed H$_3$S: inter-sublattice Coulomb coupling in a high-$\textit{T}$$_C$ superconductor | Upon thermal annealing at or above room temperature (RT) and high pressure
$\it P$ $\sim$ 155 GPa, H$_3$S exhibits superconductivity at $\it T_C$ $\sim$
200 K. Various theoretical frameworks with strong electron-phonon coupling and
Coulomb repulsion have reproduced this record-level $\it T_C$. Of particular
relevance is that observed H-D isotopic correlations among $\it T_C$, $\it P$,
and annealed order indicate limitations on the H-D isotope effect, leaving open
for consideration unconventional high-$\it T_C$ superconductivity with
electronic-based enhancements. The present work examines Coulombic pairing
arising from interactions between neighboring S and H species on separate
interlaced sublattices constituting H$_3$S in the Im$\overline{3}$m structure.
The optimal transition temperature is calculated from $\it{T}$$_{C0}$ =
$\it{k}$$_B$$^{-1}$$\Lambda$$\it{e}$$^2$/$\ell$$\zeta$, with $\Lambda$ =
0.007465 $\AA$, inter-sublattice S-H separation spacing $\zeta$ =
$\it{a}$$_0$/$\sqrt{2}$, interaction charge linear spacing $\ell$ =
$\it{a}$$_0$(3/$\sigma$)$^{1/2}$, average participating charge fraction
$\sigma$ = 3.43 $\pm$ 0.10 estimated from theory, and lattice parameter
$\it{a}$$_0$ = 3.0823 \AA. The result $\it{T}$$_{C0}$ = 198.5 $\pm$ 3.0 K is in
excellent agreement with transition temperatures determined from resistivity
and susceptibility data. Analysis of mid-infrared reflectivity confirms
correlation between boson energy and $\zeta$$^{-1}$. Suppression of $\it T_C$
with increasing residual resistance for $<$ RT annealing is treated by
scattering-induced pair breaking. Correspondence with layered high-$\it T_C$
superconductor structures are discussed. A model considering Compton scattering
of virtual photons of energies $\leq$ $\it e$$^2$/$\zeta$ by inter-sublattice
electrons is introduced, illustrating $\Lambda$ is proportional to the reduced
electron Compton wavelength.
| 0 | 1 | 0 | 0 | 0 | 0 |
A physiology--based parametric imaging method for FDG--PET data | Parametric imaging is a compartmental approach that processes nuclear imaging
data to estimate the spatial distribution of the kinetic parameters governing
tracer flow. The present paper proposes a novel and efficient computational
method for parametric imaging which is potentially applicable to several
compartmental models of diverse complexity and which is effective in the
determination of the parametric maps of all kinetic coefficients. We consider
applications to [{18}F]-fluorodeoxyglucose Positron Emission Tomography
(FDG-PET) data and analyze the two-compartment catenary model describing the
standard FDG metabolization by an homogeneous tissue and the three-compartment
non-catenary model representing the renal physiology. We show uniqueness
theorems for both models. The proposed imaging method starts from the
reconstructed FDG-PET images of tracer concentration and preliminarily applies
image processing algorithms for noise reduction and image segmentation. The
optimization procedure solves pixelwise the non-linear inverse problem of
determining the kinetic parameters from dynamic concentration data through a
regularized Gauss-Newton iterative algorithm. The reliability of the method is
validated against synthetic data, for the two-compartment system, and
experimental real data of murine models, for the renal three-compartment
system.
| 0 | 1 | 1 | 0 | 0 | 0 |
Item Recommendation with Evolving User Preferences and Experience | Current recommender systems exploit user and item similarities by
collaborative filtering. Some advanced methods also consider the temporal
evolution of item ratings as a global background process. However, all prior
methods disregard the individual evolution of a user's experience level and how
this is expressed in the user's writing in a review community. In this paper,
we model the joint evolution of user experience, interest in specific item
facets, writing style, and rating behavior. This way we can generate individual
recommendations that take into account the user's maturity level (e.g.,
recommending art movies rather than blockbusters for a cinematography expert).
As only item ratings and review texts are observables, we capture the user's
experience and interests in a latent model learned from her reviews, vocabulary
and writing style. We develop a generative HMM-LDA model to trace user
evolution, where the Hidden Markov Model (HMM) traces her latent experience
progressing over time -- with solely user reviews and ratings as observables
over time. The facets of a user's interest are drawn from a Latent Dirichlet
Allocation (LDA) model derived from her reviews, as a function of her (again
latent) experience level. In experiments with five real-world datasets, we show
that our model improves the rating prediction over state-of-the-art baselines,
by a substantial margin. We also show, in a use-case study, that our model
performs well in the assessment of user experience levels.
| 1 | 0 | 0 | 1 | 0 | 0 |
On some integrals of Hardy | We consider some properties of integrals considered by Hardy and Koshliakov,
and which have also been further extended recently by Dixit. We establish a new
general integral formula from some observations about the digamma function. We
also obtain lower and upper bounds for Hardy's integral through properties of
the digamma function.
| 0 | 0 | 1 | 0 | 0 | 0 |
Complexity of short generating functions | We give complexity analysis of the class of short generating functions (GF).
Assuming $\#P \not\subseteq FP/poly$, we show that this class is not closed
under taking many intersections, unions or projections of GFs, in the sense
that these operations can increase the bitlength of coefficients of GFs by a
super-polynomial factor. We also prove that truncated theta functions are hard
in this class.
| 1 | 0 | 1 | 0 | 0 | 0 |
Finite-Size Effects in Non-Neutral Two-Dimensional Coulomb Fluids | Thermodynamic potential of a neutral two-dimensional (2D) Cou\-lomb fluid,
confined to a large domain with a smooth boundary, exhibits at any (inverse)
temperature $\beta$ a logarithmic finite-size correction term whose universal
prefactor depends only on the Euler number of the domain and the conformal
anomaly number $c=-1$. A minimal free boson conformal field theory, which is
equivalent to the 2D symmetric two-component plasma of elementary $\pm e$
charges at coupling constant $\Gamma=\beta e^2$, was studied in the past. It
was shown that creating a non-neutrality by spreading out a charge $Q e$ at
infinity modifies the anomaly number to $c(Q,\Gamma) = - 1 + 3\Gamma Q^2$.
Here, we study the effect of non-neutrality on the finite-size expansion of the
free energy for another Coulomb fluid, namely the 2D one-component plasma
(jellium) composed of identical pointlike $e$-charges in a homogeneous
background surface charge density. For the disk geometry of the confining
domain we find that the non-neutrality induces the same change of the anomaly
number in the finite-size expansion. We derive this result first at the
free-fermion coupling $\Gamma\equiv\beta e^2=2$ and then, by using a mapping of
the 2D one-component plasma onto an anticommuting field theory formulated on a
chain, for an arbitrary coupling constant.
| 0 | 1 | 0 | 0 | 0 | 0 |
Type Safe Redis Queries: A Case Study of Type-Level Programming in Haskell | Redis is an in-memory data structure store, often used as a database, with a
Haskell interface Hedis. Redis is dynamically typed --- a key can be discarded
and re-associated to a value of a different type, and a command, when fetching
a value of a type it does not expect, signals a runtime error. We develop a
domain-specific language that, by exploiting Haskell type-level programming
techniques including indexed monad, type-level literals and closed type
families, keeps track of types of values in the database and statically
guarantees that type errors cannot happen for a class of Redis programs.
| 1 | 0 | 0 | 0 | 0 | 0 |
Ground state solutions for a nonlinear Choquard equation | We discuss the existence of ground state solutions for the Choquard equation
$$-\Delta u=(I_\alpha*F(u))F'(u)\quad\quad\quad\text{in }\mathbb R^N.$$ We
prove the existence of solutions under general hypotheses, investigating in
particular the case of a homogeneous nonlinearity $F(u)=\frac{|u|^p}p$. The
cases $N=2$ and $N\ge3$ are treated differently in some steps. The solutions
are found through a variational mountain pass strategy. The result presented
are contained in the papers with arXiv ID 1212.2027 and 1604.03294
| 0 | 0 | 1 | 0 | 0 | 0 |
Ultrashort dark solitons interactions and nonlinear tunneling in the modified nonlinear Schrödinger equation with variable coefficients | We present the study of the dark soliton dynamics in an inhomogenous fiber by
means of a variable coefficient modified nonlinear Schrödinger equation
(Vc-MNLSE) with distributed dispersion, self-phase modulation, self-steepening
and linear gain/loss. The ultrashort dark soliton pulse evolution and
interaction is studied by using the Hirota bilinear (HB) method. In particular,
we give much insight into the effect of self-steepening (SS) on the dark
soliton dynamics. The study reveals a shock wave formation, as a major effect
of SS. Numerically, we study the dark soliton propagation in the continuous
wave background, and the stability of the soliton solution is tested in the
presence of photon noise. The elastic collision behaviors of the dark solitons
are discussed by the asymptotic analysis. On the other hand, considering the
nonlinear tunneling of dark soliton through barrier/well, we find that the
tunneling of the dark soliton depends on the height of the barrier and the
amplitude of the soliton. The intensity of the tunneling soliton either forms a
peak or valley and retains its shape after the tunneling. For the case of
exponential background, the soliton tends to compress after tunneling through
the barrier/well.
| 0 | 1 | 0 | 0 | 0 | 0 |
Long-range p-d exchange interaction in a ferromagnet-semiconductor Co/CdMgTe/CdTe quantum well hybrid structure | The exchange interaction between magnetic ions and charge carriers in
semiconductors is considered as prime tool for spin control. Here, we solve a
long-standing problem by uniquely determining the magnitude of the long-range
$p-d$ exchange interaction in a ferromagnet-semiconductor (FM-SC) hybrid
structure where a 10~nm thick CdTe quantum well is separated from the FM Co
layer by a CdMgTe barrier with a thickness on the order of 10~nm. The exchange
interaction is manifested by the spin splitting of acceptor bound holes in the
effective magnetic field induced by the FM. The exchange splitting is directly
evaluated using spin-flip Raman scattering by analyzing the dependence of the
Stokes shift $\Delta_S$ on the external magnetic field $B$. We show that in
strong magnetic field $\Delta_S$ is a linear function of $B$ with an offset of
$\Delta_{pd} = 50-100~\mu$eV at zero field from the FM induced effective
exchange field. On the other hand, the $s-d$ exchange interaction between
conduction band electrons and FM, as well as the $p-d$ contribution for free
valence band holes, are negligible. The results are well described by the model
of indirect exchange interaction between acceptor bound holes in the CdTe
quantum well and the FM layer mediated by elliptically polarized phonons in the
hybrid structure.
| 0 | 1 | 0 | 0 | 0 | 0 |
Parsec-scale Faraday rotation and polarization of 20 active galactic nuclei jets | We perform polarimetry analysis of 20 active galactic nuclei (AGN) jets using
the Very Long Baseline Array (VLBA) at 1.4, 1.6, 2.2, 2.4, 4.6, 5.0, 8.1, 8.4,
and 15.4 GHz. The study allowed us to investigate linearly polarized properties
of the jets at parsec-scales: distribution of the Faraday rotation measure (RM)
and fractional polarization along the jets, Faraday effects and structure of
Faraday-corrected polarization images. Wavelength-dependence of the fractional
polarization and polarization angle is consistent with external Faraday
rotation, while some sources show internal rotation. The RM changes along the
jets, systematically increasing its value towards synchrotron self-absorbed
cores at shorter wavelengths. The highest core RM reaches 16,900 rad/m^2 in the
source rest frame for the quasar 0952+179, suggesting the presence of highly
magnetized, dense media in these regions. The typical RM of transparent jet
regions has values of an order of a hundred rad/m^2. Significant transverse
rotation measure gradients are observed in seven sources. The magnetic field in
the Faraday screen has no preferred orientation, and is observed to be random
or regular from source to source. Half of the sources show evidence for the
helical magnetic fields in their rotating magnetoionic media. At the same time
jets themselves contain large-scale, ordered magnetic fields and tend to align
its direction with the jet flow. The observed variety of polarized signatures
can be explained by a model of spine-sheath jet structure.
| 0 | 1 | 0 | 0 | 0 | 0 |
Positive scalar curvature and connected sums | Let $N$ be a closed enlargeable manifold in the sense of Gromov-Lawson and
$M$ a closed spin manifold of equal dimension, a famous theorem of
Gromov-Lawson states that the connected sum $M\# N$ admits no metric of
positive scalar curvature. We present a potential generalization of this result
to the case where $M$ is nonspin. We use index theory for Dirac operators to
prove our result.
| 0 | 0 | 1 | 0 | 0 | 0 |
Optical computing by injection-locked lasers | A programmable optical computer has remained an elusive concept. To construct
a practical computing primitive equivalent to an electronic Boolean logic, one
should find a nonlinear phenomenon that overcomes weaknesses present in many
optical processing schemes. Ideally, the nonlinearity should provide a
functionally complete set of logic operations, enable ultrafast all-optical
programmability, and allow cascaded operations without a change in the
operating wavelength or in the signal encoding format. Here we demonstrate a
programmable logic gate using an injection-locked Vertical-Cavity
Surface-Emitting Laser (VCSEL). The gate program is switched between the AND
and the OR operations at the rate of 1 GHz with Bit Error Ratio (BER) of 10e-6
without changes in the wavelength or in the signal encoding format. The scheme
is based on nonlinearity of normalization operations, which can be used to
construct any continuous complex function or operation, Boolean or otherwise.
| 1 | 0 | 0 | 0 | 0 | 0 |
On the joint asymptotic distribution of the restricted estimators in multivariate regression model | The main Theorem of Jain et al.[Jain, K., Singh, S., and Sharma, S. (2011),
Re- stricted estimation in multivariate measurement error regression model;
JMVA, 102, 2, 264-280] is established in its full generality. Namely, we derive
the joint asymp- totic normality of the unrestricted estimator (UE) and the
restricted estimators of the matrix of the regression coefficients. The derived
result holds under the hypothesized restriction as well as under the sequence
of alternative restrictions. In addition, we establish Asymptotic
Distributional Risk for the estimators and compare their relative performance.
It is established that near the restriction, the restricted estimators (REs)
perform better than the UE. But the REs perform worse than the unrestricted
estimator when one moves far away from the restriction.
| 0 | 0 | 1 | 1 | 0 | 0 |
Rotational spectroscopy, tentative interstellar detection, and chemical modelling of N-methylformamide | N-methylformamide, CH3NHCHO, may be an important molecule for interstellar
pre-biotic chemistry because it contains a peptide bond. The rotational
spectrum of the most stable trans conformer of CH3NHCHO is complicated by
strong torsion-rotation interaction due to the low barrier of the methyl
torsion. We use two absorption spectrometers in Kharkiv and Lille to measure
the rotational spectra over 45--630 GHz. The analysis is carried out using the
Rho-axis method and the RAM36 code. We search for N-methylformamide toward the
hot molecular core Sgr B2(N2) using a spectral line survey carried out with
ALMA. The astronomical results are put into a broader astrochemical context
with the help of a gas-grain chemical kinetics model. The laboratory data set
for the trans conformer of CH3NHCHO consists of 9469 line frequencies with J <=
62, including the first assignment of the rotational spectra of the first and
second excited torsional states. All these lines are fitted within experimental
accuracy. We report the tentative detection of CH3NHCHO towards Sgr B2(N2). We
find CH3NHCHO to be more than one order of magnitude less abundant than NH2CHO,
a factor of two less abundant than CH3NCO, but only slightly less abundant than
CH3CONH2. The chemical models indicate that the efficient formation of HNCO via
NH + CO on grains is a necessary step in the achievement of the observed
gas-phase abundance of CH3NCO. Production of CH3NHCHO may plausibly occur on
grains either through the direct addition of functional-group radicals or
through the hydrogenation of CH3NCO. Provided the detection of CH3NHCHO is
confirmed, the only slight underabundance of this molecule compared to its more
stable structural isomer acetamide and the sensitivity of the model abundances
to the chemical kinetics parameters suggest that the formation of these two
molecules is controlled by kinetics rather than thermal equilibrium.
| 0 | 1 | 0 | 0 | 0 | 0 |
Dynamic anisotropy in MHD turbulence induced by mean magnetic field | In this paper, we study the development of anisotropy in strong MHD
turbulence in the presence of a large scale magnetic field B 0 by analyzing the
results of direct numerical simulations. Our results show that the developed
anisotropy among the different components of the velocity and magnetic field is
a direct outcome of the inverse cascade of energy of the perpendicular velocity
components u? and a forward cascade of the energy of the parallel component u k
. The inverse cascade develops for a strong B0, where the flow exhibits a
strong vortical structure by the suppression of fluctuations along the magnetic
field. Both the inverse and the forward cascade are examined in detail by
investigating the anisotropic energy spectra, the energy fluxes, and the shell
to shell energy transfers among different scales.
| 0 | 1 | 0 | 0 | 0 | 0 |
Dealing with the exponential wall in electronic structure calculations | An alternative to Density Functional Theory are wavefunction based electronic
structure calculations for solids. In order to perform them the Exponential
Wall (EW) problem has to be resolved. It is caused by an exponential increase
of the number of configurations with increasing electron number N. There are
different routes one may follow. One is to characterize a many-electron
wavefunction by a vector in Liouville space with a cumulant metric rather than
in Hilbert space. This removes the EW problem. Another is to model the solid by
an {\it impurity} or {\it fragment} embedded in a {\it bath} which is treated
at a much lower level than the former. This is the case in Density Matrix
Embedding Theory (DMET) or Density Embedding Theory (DET). The latter are
closely related to a Schmidt decomposition of a system and to the determination
of the associated entanglement. We show here the connection between the two
approaches. It turns out that the DMET (or DET) has an identical active space
as a previously used Local Ansatz, based on a projection and partitioning
approach. Yet, the EW problem is resolved differently in the two cases. By
studying a $H_{10}$ ring these differences are analyzed with the help of the
method of increments.
| 0 | 1 | 0 | 0 | 0 | 0 |
Influence maximization on correlated networks through community identification | The identification of the minimal set of nodes that maximizes the propagation
of information is one of the most important problems in network science. In
this paper, we introduce a new method to find the set of initial spreaders to
maximize the information propagation in complex networks. We evaluate this
method in assortative networks and verify that degree-degree correlation plays
a fundamental role on the spreading dynamics. Simulation results show that our
algorithm is statistically similar, in terms of the average size of outbreaks,
to the greedy approach. However, our method is much less time consuming than
the greedy algorithm.
| 1 | 1 | 0 | 0 | 0 | 0 |
Conformal metrics with prescribed fractional scalar curvature on conformal infinities with positive fractional Yamabe constants | Let $(X, g^+)$ be an asymptotically hyperbolic manifold and $(M, [\hat{h}])$
its conformal infinity. Our primary aim in this paper is to introduce the
prescribed fractional scalar curvature problem on $M$ and provide solutions
under various geometric conditions on $X$ and $M$. We also obtain the existence
results for the fractional Yamabe problem in the endpoint case, e.g., $n = 3$,
$\gamma = 1/2$ and $M$ is non-umbilic, etc. Every solution we find turns out to
be smooth on $M$.
| 0 | 0 | 1 | 0 | 0 | 0 |
A Short Review of Ethical Challenges in Clinical Natural Language Processing | Clinical NLP has an immense potential in contributing to how clinical
practice will be revolutionized by the advent of large scale processing of
clinical records. However, this potential has remained largely untapped due to
slow progress primarily caused by strict data access policies for researchers.
In this paper, we discuss the concern for privacy and the measures it entails.
We also suggest sources of less sensitive data. Finally, we draw attention to
biases that can compromise the validity of empirical research and lead to
socially harmful applications.
| 1 | 0 | 0 | 0 | 0 | 0 |
kd-switch: A Universal Online Predictor with an application to Sequential Two-Sample Testing | We propose a novel online predictor for discrete labels conditioned on
multivariate features in $\mathbb{R}^d$. The predictor is pointwise universal:
it achieves a normalized log loss performance asymptotically as good as the
true conditional entropy of the labels given the features. The predictor is
based on a feature space discretization induced by a full-fledged k-d tree with
randomly picked directions and a switch distribution, requiring no
hyperparameter setting and automatically selecting the most relevant scales in
the feature space. Using recent results, a consistent sequential two-sample
test is built from this predictor. In terms of discrimination power, on
selected challenging datasets, it is comparable to or better than
state-of-the-art non-sequential two-sample tests based on the train-test
paradigm and, a recent sequential test requiring hyperparameters. The time
complexity to process the $n$-th sample point is $O(\log n)$ in probability
(with respect to the distribution generating the data points), in contrast to
the linear complexity of the previous sequential approach.
| 1 | 0 | 0 | 1 | 0 | 0 |
A Heat Equation on some Adic Completions of Q and Ultrametric Analysis | This article deals with a Markov process related to the fundamental solution
of a heat equation on the direct product ring Q_S, where Q_S is a finite direct
product of p-adic fields. The techniques developed here are different from the
well known ones: they are geometrical and very simple. As a result, the
techniques developed here provides a general framework of these problems on
other related ultrametric groups.
| 0 | 0 | 1 | 0 | 0 | 0 |
A simple alteration of the peridynamics correspondence principle to eliminate zero-energy deformation | We look for an enhancement of the correspondence model of peridynamics with a
view to eliminating the zero-energy deformation modes. Since the non-local
integral definition of the deformation gradient underlies the problem, we
initially look for a remedy by introducing a class of localizing corrections to
the integral. Since the strategy is found to afford only a reduction, and not
complete elimination, of the oscillatory zero-energy deformation, we propose in
the sequel an alternative approach based on the notion of sub-horizons. A most
useful feature of the last proposal is that the setup, whilst providing the
solution with the necessary stability, deviates only marginally from the
original correspondence formulation. We also undertake a set of numerical
simulations that attest to the remarkable efficacy of the sub-horizon based
methodology.
| 0 | 1 | 0 | 0 | 0 | 0 |
Interpretable Active Learning | Active learning has long been a topic of study in machine learning. However,
as increasingly complex and opaque models have become standard practice, the
process of active learning, too, has become more opaque. There has been little
investigation into interpreting what specific trends and patterns an active
learning strategy may be exploring. This work expands on the Local
Interpretable Model-agnostic Explanations framework (LIME) to provide
explanations for active learning recommendations. We demonstrate how LIME can
be used to generate locally faithful explanations for an active learning
strategy, and how these explanations can be used to understand how different
models and datasets explore a problem space over time. In order to quantify the
per-subgroup differences in how an active learning strategy queries spatial
regions, we introduce a notion of uncertainty bias (based on disparate impact)
to measure the discrepancy in the confidence for a model's predictions between
one subgroup and another. Using the uncertainty bias measure, we show that our
query explanations accurately reflect the subgroup focus of the active learning
queries, allowing for an interpretable explanation of what is being learned as
points with similar sources of uncertainty have their uncertainty bias
resolved. We demonstrate that this technique can be applied to track
uncertainty bias over user-defined clusters or automatically generated clusters
based on the source of uncertainty.
| 1 | 0 | 0 | 1 | 0 | 0 |
Orbital-Free Density-Functional Theory Simulations of Displacement Cascade in Aluminum | Here, we report orbital-free density-functional theory (OF DFT) molecular
dynamics simulations of the displacement cascade in aluminum. The electronic
effect is our main concern. The displacement threshold energies are calculated
using OF DFT and classical molecular dynamics (MD) and the comparison reveals
the role of charge bridge. Compared to MD simulation, the displacement spike
from OF DFT has a lower peak and shorter duration time, which is attributed to
the effect of electronic damping. The charge density profiles clearly display
the existence of depleted zones, vacancy and interstitial clusters. And it is
found that the energy exchanges between ions and electrons are mainly
contributed by the kinetic energies.
| 0 | 1 | 0 | 0 | 0 | 0 |
Binary systems with an RR Lyrae component - progress in 2016 | In this contribution, we summarize the progress made in the investigation of
binary candidates with an RR Lyrae component in 2016. We also discuss the
actual status of the RRLyrBinCan database.
| 0 | 1 | 0 | 0 | 0 | 0 |
Covering and tiling hypergraphs with tight cycles | Given $3 \leq k \leq s$, we say that a $k$-uniform hypergraph $C^k_s$ is a
tight cycle on $s$ vertices if there is a cyclic ordering of the vertices of
$C^k_s$ such that every $k$ consecutive vertices under this ordering form an
edge. We prove that if $k \ge 3$ and $s \ge 2k^2$, then every $k$-uniform
hypergraph on $n$ vertices with minimum codegree at least $(1/2 + o(1))n$ has
the property that every vertex is covered by a copy of $C^k_s$. Our result is
asymptotically best possible for infinitely many pairs of $s$ and $k$, e.g.
when $s$ and $k$ are coprime.
A perfect $C^k_s$-tiling is a spanning collection of vertex-disjoint copies
of $C^k_s$. When $s$ is divisible by $k$, the problem of determining the
minimum codegree that guarantees a perfect $C^k_s$-tiling was solved by a
result of Mycroft. We prove that if $k \ge 3$ and $s \ge 5k^2$ is not divisible
by $k$ and $s$ divides $n$, then every $k$-uniform hypergraph on $n$ vertices
with minimum codegree at least $(1/2 + 1/(2s) + o(1))n$ has a perfect
$C^k_s$-tiling. Again our result is asymptotically best possible for infinitely
many pairs of $s$ and $k$, e.g. when $s$ and $k$ are coprime with $k$ even.
| 0 | 0 | 1 | 0 | 0 | 0 |
DNA methylation markers to assess biological age | Among the different biomarkers of aging based on omics and clinical data, DNA
methylation clocks stand apart providing unmatched accuracy in assessing the
biological age of both humans and animal models of aging. Here, we discuss
robustness of DNA methylation clocks and bounds on their out-of-sample
performance and review computational strategies for development of the clocks.
| 0 | 0 | 0 | 0 | 1 | 0 |
Towards Building an Intelligent Anti-Malware System: A Deep Learning Approach using Support Vector Machine (SVM) for Malware Classification | Effective and efficient mitigation of malware is a long-time endeavor in the
information security community. The development of an anti-malware system that
can counteract an unknown malware is a prolific activity that may benefit
several sectors. We envision an intelligent anti-malware system that utilizes
the power of deep learning (DL) models. Using such models would enable the
detection of newly-released malware through mathematical generalization. That
is, finding the relationship between a given malware $x$ and its corresponding
malware family $y$, $f: x \mapsto y$. To accomplish this feat, we used the
Malimg dataset (Nataraj et al., 2011) which consists of malware images that
were processed from malware binaries, and then we trained the following DL
models 1 to classify each malware family: CNN-SVM (Tang, 2013), GRU-SVM
(Agarap, 2017), and MLP-SVM. Empirical evidence has shown that the GRU-SVM
stands out among the DL models with a predictive accuracy of ~84.92%. This
stands to reason for the mentioned model had the relatively most sophisticated
architecture design among the presented models. The exploration of an even more
optimal DL-SVM model is the next stage towards the engineering of an
intelligent anti-malware system.
| 0 | 0 | 0 | 1 | 0 | 0 |
Exponential profiles from stellar scattering off interstellar clumps and holes in dwarf galaxy discs | Holes and clumps in the interstellar gas of dwarf irregular galaxies are
gravitational scattering centers that heat field stars and change their radial
and vertical distributions. Because the gas structures are extended and each
stellar scattering is relatively weak, the stellar orbits remain nearly
circular and the net effect accumulates slowly over time. We calculate the
radial profile of scattered stars with an idealized model and find that it
approaches an equilibrium shape that is exponential, similar to the observed
shapes of galaxy discs. Our models treat only scattering and have no bars or
spiral arms, so the results apply mostly to dwarf irregular galaxies where
there are no other obvious scattering processes. Stellar scattering by gaseous
perturbations slows down when the stellar population gets thicker than the gas
layer. An accreting galaxy with a growing thin gas layer can form multiple
stellar exponential profiles from the inside-out, preserving the remnants of
each Gyr interval in a sequence of ever-lengthening and thinning stellar
subdiscs.
| 0 | 1 | 0 | 0 | 0 | 0 |
Visualizations for an Explainable Planning Agent | In this paper, we report on the visualization capabilities of an Explainable
AI Planning (XAIP) agent that can support human in the loop decision making.
Imposing transparency and explainability requirements on such agents is
especially important in order to establish trust and common ground with the
end-to-end automated planning system. Visualizing the agent's internal
decision-making processes is a crucial step towards achieving this. This may
include externalizing the "brain" of the agent -- starting from its sensory
inputs, to progressively higher order decisions made by it in order to drive
its planning components. We also show how the planner can bootstrap on the
latest techniques in explainable planning to cast plan visualization as a plan
explanation problem, and thus provide concise model-based visualization of its
plans. We demonstrate these functionalities in the context of the automated
planning components of a smart assistant in an instrumented meeting space.
| 1 | 0 | 0 | 0 | 0 | 0 |
Generative adversarial interpolative autoencoding: adversarial training on latent space interpolations encourage convex latent distributions | We present a neural network architecture based upon the Autoencoder (AE) and
Generative Adversarial Network (GAN) that promotes a convex latent distribution
by training adversarially on latent space interpolations. By using an AE as
both the generator and discriminator of a GAN, we pass a pixel-wise error
function across the discriminator, yielding an AE which produces non-blurry
samples that match both high- and low-level features of the original images.
Interpolations between images in this space remain within the latent-space
distribution of real images as trained by the discriminator, and therfore
preserve realistic resemblances to the network inputs.
| 0 | 0 | 0 | 1 | 0 | 0 |
Attribute-Guided Face Generation Using Conditional CycleGAN | We are interested in attribute-guided face generation: given a low-res face
input image, an attribute vector that can be extracted from a high-res image
(attribute image), our new method generates a high-res face image for the
low-res input that satisfies the given attributes. To address this problem, we
condition the CycleGAN and propose conditional CycleGAN, which is designed to
1) handle unpaired training data because the training low/high-res and high-res
attribute images may not necessarily align with each other, and to 2) allow
easy control of the appearance of the generated face via the input attributes.
We demonstrate impressive results on the attribute-guided conditional CycleGAN,
which can synthesize realistic face images with appearance easily controlled by
user-supplied attributes (e.g., gender, makeup, hair color, eyeglasses). Using
the attribute image as identity to produce the corresponding conditional vector
and by incorporating a face verification network, the attribute-guided network
becomes the identity-guided conditional CycleGAN which produces impressive and
interesting results on identity transfer. We demonstrate three applications on
identity-guided conditional CycleGAN: identity-preserving face superresolution,
face swapping, and frontal face generation, which consistently show the
advantage of our new method.
| 1 | 0 | 0 | 1 | 0 | 0 |
Objective-Reinforced Generative Adversarial Networks (ORGAN) for Sequence Generation Models | In unsupervised data generation tasks, besides the generation of a sample
based on previous observations, one would often like to give hints to the model
in order to bias the generation towards desirable metrics. We propose a method
that combines Generative Adversarial Networks (GANs) and reinforcement learning
(RL) in order to accomplish exactly that. While RL biases the data generation
process towards arbitrary metrics, the GAN component of the reward function
ensures that the model still remembers information learned from data. We build
upon previous results that incorporated GANs and RL in order to generate
sequence data and test this model in several settings for the generation of
molecules encoded as text sequences (SMILES) and in the context of music
generation, showing for each case that we can effectively bias the generation
process towards desired metrics.
| 1 | 0 | 0 | 1 | 0 | 0 |
Pluripotential theory on the support of closed positive currents and applications to dynamics in $\mathbb{C}^n$ | We extend certain classical theorems in pluripotential theory to a class of
functions defined on the support of a $(1,1)$-closed positive current $T$,
analogous to plurisubharmonic functions, called $T$-plurisubharmonic functions.
These functions are defined as limits, on the support of $T$, of sequences of
plurisubharmonic functions decreasing on this support. In particular, we show
that the poles of such functions are pluripolar sets. We also show that the
maximum principle and the Hartogs's theorem remain valid in a weak sense. We
study these functions by means of a class of measures, so-called "pluri-Jensen
measures", about which we prove that they are numerous on the support of
$(1,1)$-closed positive currents. We also obtain, for any fat compact set, an
expression of its relative Green's function in terms of an infimum of an
integral over a set of pluri-Jensen measures. We then deduce, by means of these
measures, a characterization of the polynomially convex fat compact sets, as
well as a characterization of pluripolar sets, and the fact that the support of
a closed positive $(1,1)$-current is nowhere pluri-thin. In the second part of
this article, these tools are used to study dynamics of a certain class of
automorphisms of $\mathbb{C}^n$ which naturally generalize Hénon's
automorphisms of $\mathbb{C}^2$. First we study the geometry of the support of
canonical invariant currents. Then we obtain an equidistribution result for the
convergence of pull-back of certain measures towards an ergodic invariant
measure, with compact support.
| 0 | 0 | 1 | 0 | 0 | 0 |
Inconsistency of Measure-Theoretic Probability and Random Behavior of Microscopic Systems | We report an inconsistency found in probability theory (also referred to as
measure-theoretic probability). For probability measures induced by real-valued
random variables, we deduce an "equality" such that one side of the "equality"
is a probability, but the other side is not. For probability measures induced
by extended random variables, we deduce an "equality" such that its two sides
are unequal probabilities. The deduced expressions are erroneous only when it
can be proved that measure-theoretic probability is a theory free from
contradiction. However, such a proof does not exist. The inconsistency appears
only in the theory rather than in the physical world, and will not affect
practical applications as long as ideal events in the theory (which will not
occur physically) are not mistaken for observable events in the real world.
Nevertheless, unlike known paradoxes in mathematics, the inconsistency cannot
be explained away and hence must be resolved. The assumption of infinite
additivity in the theory is relevant to the inconsistency, and may cause
confusion of ideal events and real events. As illustrated by an example in this
article, since abstract properties of mathematical entities in theoretical
thinking are not necessarily properties of physical quantities observed in the
real world, mistaking the former for the latter may lead to misinterpreting
random phenomena observed in experiments with microscopic systems. Actually the
inconsistency is due to the notion of "numbers" adopted in conventional
mathematics. A possible way to resolve the inconsistency is to treat "numbers"
from the viewpoint of constructive mathematics.
| 0 | 0 | 1 | 0 | 0 | 0 |
Robustness of Quasiparticle Interference Test for Sign-changing Gaps in Multiband Superconductors | Recently, a test for a sign-changing gap function in a candidate multiband
unconventional superconductor involving quasiparticle interference data was
proposed. The test was based on the antisymmetric, Fourier transformed
conductance maps integrated over a range of momenta $\bf q$ corresponding to
interband processes, which was argued to display a particular resonant form,
provided the gaps changed sign between the Fermi surface sheets connected by
$\bf q$. The calculation was performed for a single impurity, however, raising
the question of how robust this measure is as a test of sign-changing pairing
in a realistic system with many impurities. Here we reproduce the results of
the previous work within a model with two distinct Fermi surface sheets, and
show explicitly that the previous result, while exact for a single nonmagnetic
scatterer and also in the limit of a dense set of random impurities, can be
difficult to implement for a few dilute impurities. In this case, however,
appropriate isolation of a single impurity is sufficient to recover the
expected result, allowing a robust statement about the gap signs to be made.
| 0 | 1 | 0 | 0 | 0 | 0 |
Adaptive User Perspective Rendering for Handheld Augmented Reality | Handheld Augmented Reality commonly implements some variant of magic lens
rendering, which turns only a fraction of the user's real environment into AR
while the rest of the environment remains unaffected. Since handheld AR devices
are commonly equipped with video see-through capabilities, AR magic lens
applications often suffer from spatial distortions, because the AR environment
is presented from the perspective of the camera of the mobile device. Recent
approaches counteract this distortion based on estimations of the user's head
position, rendering the scene from the user's perspective. To this end,
approaches usually apply face-tracking algorithms on the front camera of the
mobile device. However, this demands high computational resources and therefore
commonly affects the performance of the application beyond the already high
computational load of AR applications. In this paper, we present a method to
reduce the computational demands for user perspective rendering by applying
lightweight optical flow tracking and an estimation of the user's motion before
head tracking is started. We demonstrate the suitability of our approach for
computationally limited mobile devices and we compare it to device perspective
rendering, to head tracked user perspective rendering, as well as to fixed
point of view user perspective rendering.
| 1 | 0 | 0 | 0 | 0 | 0 |
A distributed primal-dual algorithm for computation of generalized Nash equilibria with shared affine coupling constraints via operator splitting methods | In this paper, we propose a distributed primal-dual algorithm for computation
of a generalized Nash equilibrium (GNE) in noncooperative games over network
systems. In the considered game, not only each player's local objective
function depends on other players' decisions, but also the feasible decision
sets of all the players are coupled together with a globally shared affine
inequality constraint. Adopting the variational GNE, that is the solution of a
variational inequality, as a refinement of GNE, we introduce a primal-dual
algorithm that players can use to seek it in a distributed manner. Each player
only needs to know its local objective function, local feasible set, and a
local block of the affine constraint. Meanwhile, each player only needs to
observe the decisions on which its local objective function explicitly depends
through the interference graph and share information related to multipliers
with its neighbors through a multiplier graph. Through a primal-dual analysis
and an augmentation of variables, we reformulate the problem as finding the
zeros of a sum of monotone operators. Our distributed primal-dual algorithm is
based on forward-backward operator splitting methods. We prove its convergence
to the variational GNE for fixed step-sizes under some mild assumptions. Then a
distributed algorithm with inertia is also introduced and analyzed for
variational GNE seeking. Finally, numerical simulations for network Cournot
competition are given to illustrate the algorithm efficiency and performance.
| 1 | 0 | 1 | 0 | 0 | 0 |
Inverse dynamic and spectral problems for the one-dimensional Dirac system on a finite tree | We consider inverse dynamic and spectral problems for the one dimensional
Dirac system on a finite tree. Our aim will be to recover the topology of a
tree (lengths and connectivity of edges) as well as matrix potentials on each
edge. As inverse data we use the Weyl-Titchmarsh matrix function or the dynamic
response operator.
| 0 | 0 | 1 | 0 | 0 | 0 |
Extensions of isomorphisms of subvarieties in flexile varieties | Let $X$ be a quasi-affine algebraic variety isomorphic to the complement of a
closed subvariety of dimension at most $n-3$ in $\C^n$. We find some conditions
under which an isomorphism of two closed subvarieties of $X$ can be extended to
an automorphism of $X$.
| 0 | 0 | 1 | 0 | 0 | 0 |
Frequentist coverage and sup-norm convergence rate in Gaussian process regression | Gaussian process (GP) regression is a powerful interpolation technique due to
its flexibility in capturing non-linearity. In this paper, we provide a general
framework for understanding the frequentist coverage of point-wise and
simultaneous Bayesian credible sets in GP regression. As an intermediate
result, we develop a Bernstein von-Mises type result under supremum norm in
random design GP regression. Identifying both the mean and covariance function
of the posterior distribution of the Gaussian process as regularized
$M$-estimators, we show that the sampling distribution of the posterior mean
function and the centered posterior distribution can be respectively
approximated by two population level GPs. By developing a comparison inequality
between two GPs, we provide exact characterization of frequentist coverage
probabilities of Bayesian point-wise credible intervals and simultaneous
credible bands of the regression function. Our results show that inference
based on GP regression tends to be conservative; when the prior is
under-smoothed, the resulting credible intervals and bands have minimax-optimal
sizes, with their frequentist coverage converging to a non-degenerate value
between their nominal level and one. As a byproduct of our theory, we show that
the GP regression also yields minimax-optimal posterior contraction rate
relative to the supremum norm, which provides a positive evidence to the long
standing problem on optimal supremum norm contraction rate in GP regression.
| 0 | 0 | 1 | 1 | 0 | 0 |
Most Probable Evolution Trajectories in a Genetic Regulatory System Excited by Stable Lévy Noise | We study the most probable trajectories of the concentration evolution for
the transcription factor activator in a genetic regulation system, with
non-Gaussian stable Lévy noise in the synthesis reaction rate taking into
account. We calculate the most probable trajectory by spatially maximizing the
probability density of the system path, i.e., the solution of the associated
nonlocal Fokker-Planck equation. We especially examine those most probable
trajectories from low concentration state to high concentration state (i.e.,
the likely transcription regime) for certain parameters, in order to gain
insights into the transcription processes and the tipping time for the
transcription likely to occur. This enables us: (i) to visualize the progress
of concentration evolution (i.e., observe whether the system enters the
transcription regime within a given time period); (ii) to predict or avoid
certain transcriptions via selecting specific noise parameters in particular
regions in the parameter space. Moreover, we have found some peculiar or
counter-intuitive phenomena in this gene model system, including (a) a smaller
noise intensity may trigger the transcription process, while a larger noise
intensity can not, under the same asymmetric Lévy noise. This phenomenon does
not occur in the case of symmetric Lévy noise; (b) the symmetric Lévy
motion always induces transition to high concentration, but certain asymmetric
Lévy motions do not trigger the switch to transcription. These findings
provide insights for further experimental research, in order to achieve or to
avoid specific gene transcriptions, with possible relevance for medical
advances.
| 0 | 0 | 0 | 0 | 1 | 0 |
Simplified derivation of the collision probability of two objects in independent Keplerian orbits | Many topics in planetary studies demand an estimate of the collision
probability of two objects moving on nearly Keplerian orbits. In the classic
works of Öpik (1951) and Wetherill (1967), the collision probability was
derived by linearizing the motion near the collision points, and there is now a
vast literature using their method. We present here a simpler and more
physically motivated derivation for non-tangential collisions in Keplerian
orbits, as well as for tangential collisions that were not previously
considered. Our formulas have the added advantage of being manifestly symmetric
in the parameters of the two colliding bodies. In common with the
Öpik-Wetherill treatments, we linearize the motion of the bodies in the
vicinity of the point of orbit intersection (or near the points of minimum
distance between the two orbits) and assume a uniform distribution of impact
parameter within the collision radius. We point out that the linear
approximation leads to singular results for the case of tangential encounters.
We regularize this singularity by use of a parabolic approximation of the
motion in the vicinity of a tangential encounter.
| 0 | 1 | 0 | 0 | 0 | 0 |
Optically Coupled Methods for Microwave Impedance Microscopy | Scanning Microwave Impedance Microscopy (MIM) measurement of
photoconductivity with 50 nm resolution is demonstrated using a modulated
optical source. The use of a modulated source allows for measurement of
photoconductivity in a single scan without a reference region on the sample, as
well as removing most topographical artifacts and enhancing signal to noise as
compared with unmodulated measurement. A broadband light source with tunable
monochrometer is then used to measure energy resolved photoconductivity with
the same methodology. Finally, a pulsed optical source is used to measure local
photo-carrier lifetimes via MIM, using the same 50 nm resolution tip.
| 0 | 1 | 0 | 0 | 0 | 0 |
Generalized feedback vertex set problems on bounded-treewidth graphs: chordality is the key to single-exponential parameterized algorithms | It has long been known that Feedback Vertex Set can be solved in time
$2^{\mathcal{O}(w\log w)}n^{\mathcal{O}(1)}$ on $n$-vertex graphs of treewidth
$w$, but it was only recently that this running time was improved to
$2^{\mathcal{O}(w)}n^{\mathcal{O}(1)}$, that is, to single-exponential
parameterized by treewidth. We investigate which generalizations of Feedback
Vertex Set can be solved in a similar running time. Formally, for a class
$\mathcal{P}$ of graphs, the Bounded $\mathcal{P}$-Block Vertex Deletion
problem asks, given a graph~$G$ on $n$ vertices and positive integers~$k$
and~$d$, whether $G$ contains a set~$S$ of at most $k$ vertices such that each
block of $G-S$ has at most $d$ vertices and is in $\mathcal{P}$. Assuming that
$\mathcal{P}$ is recognizable in polynomial time and satisfies a certain
natural hereditary condition, we give a sharp characterization of when
single-exponential parameterized algorithms are possible for fixed values of
$d$: if $\mathcal{P}$ consists only of chordal graphs, then the problem can be
solved in time $2^{\mathcal{O}(wd^2)} n^{\mathcal{O}(1)}$, and if $\mathcal{P}$
contains a graph with an induced cycle of length $\ell\ge 4$, then the problem
is not solvable in time $2^{o(w\log w)} n^{\mathcal{O}(1)}$ even for fixed
$d=\ell$, unless the ETH fails. We also study a similar problem, called Bounded
$\mathcal{P}$-Component Vertex Deletion, where the target graphs have connected
components of small size rather than blocks of small size, and we present
analogous results. For this problem, we also show that if $d$ is part of the
input and $\mathcal{P}$ contains all chordal graphs, then it cannot be solved
in time $f(w)n^{o(w)}$ for some function $f$, unless the ETH fails.
| 1 | 0 | 0 | 0 | 0 | 0 |
Electron-phonon coupling mechanisms for hydrogen-rich metals at high pressure | The mechanisms for strong electron-phonon coupling predicted for
hydrogen-rich alloys with high superconducting critical temperature ($T_c$) are
examined within the Migdal-Eliashberg theory. Analysis of the functional
derivative of $T_c$ with respect to the electron-phonon spectral function shows
that at low pressures, when the alloys often adopt layered structures, bending
vibrations have the most dominant effect. At very high pressures, the H-H
interactions in two-dimensional (2D) and three-dimensional (3D) extended
structures are weakened, resulting in mixed bent (libration) and stretch
vibrations, and the electron-phonon coupling process is distributed over a
broad frequency range leading to very high $T_c$.
| 0 | 1 | 0 | 0 | 0 | 0 |
Competitive Equilibrium For almost All Incomes | Competitive equilibrium from equal incomes (CEEI) is a well-known rule for
fair allocation of resources among agents with different preferences. It has
many advantages, among them is the fact that a CEEI allocation is both Pareto
efficient and envy-free. However, when the resources are indivisible, a CEEI
allocation might not exist even when there are two agents and a single item.
In contrast to this discouraging non-existence result, Babaioff, Nisan and
Talgam-Cohen (2017) recently suggested a new and more encouraging approach to
allocation of indivisible items: instead of insisting that the incomes be
equal, they suggest to look at the entire space of possible incomes, and check
whether there exists a competitive equilibrium for almost all income-vectors
(CEFAI) --- all income-space except a subset of measure zero. They show that a
CEFAI exists when there are at most 3 items, or when there are 4 items and two
agents. They also show that when there are 5 items and two agents there might
not exist a CEFAI. They leave open the cases of 4 items with three or four
agents.
This paper presents a new way to implement a CEFAI, as a subgame-perfect
equilibrium of a sequential game. This new implementation allows us both to
offer much simpler solutions to the known cases (at most 3 items, and 4 items
with two agents), and to prove that a CEFAI exists even in the much more
difficult case of 4 items and three agents. Moreover, we prove that a CEFAI
might not exist with 4 items and four agents. When the items to be divided are
bads (chores), CEFAI exists for two agents with at most 4 chores, but does not
exist for two agents with 5 chores or with three agents with 3 or more chores.
Thus, this paper completes the characterization of CEFAI existence for monotone
preferences.
| 1 | 0 | 0 | 0 | 0 | 0 |
Converting of algebraic Diophantine equations to a diagonal form with the help of an integer non-orthogonal transformation, maintaining the asymptotic behavior of the number of its integer solutions | The author showed that any homogeneous algebraic Diophantine equation of the
second order can be converted to a diagonal form using an integer
non-orthogonal transformation maintaining asymptotic behavior of the number of
its integer solutions. In this paper, we consider the transformation to the
diagonal form of a wider class of algebraic second-order Diophantine equations,
and also we consider the conditions for converting higher order algebraic
Diophantine equations to this form. The author found an asymptotic estimate for
the number of integer solutions of the diagonal Thue equation of odd degree
with an amount of variables greater than two, and also he got and asymptotic
estimates of the number of integer solutions of other types of diagonal
algebraic Diophantine equations.
| 0 | 0 | 1 | 0 | 0 | 0 |
A Deep Learning Approach with an Attention Mechanism for Automatic Sleep Stage Classification | Automatic sleep staging is a challenging problem and state-of-the-art
algorithms have not yet reached satisfactory performance to be used instead of
manual scoring by a sleep technician. Much research has been done to find good
feature representations that extract the useful information to correctly
classify each epoch into the correct sleep stage. While many useful features
have been discovered, the amount of features have grown to an extent that a
feature reduction step is necessary in order to avoid the curse of
dimensionality. One reason for the need of such a large feature set is that
many features are good for discriminating only one of the sleep stages and are
less informative during other stages. This paper explores how a second feature
representation over a large set of pre-defined features can be learned using an
auto-encoder with a selective attention for the current sleep stage in the
training batch. This selective attention allows the model to learn feature
representations that focuses on the more relevant inputs without having to
perform any dimensionality reduction of the input data. The performance of the
proposed algorithm is evaluated on a large data set of polysomnography (PSG)
night recordings of patients with sleep-disordered breathing. The performance
of the auto-encoder with selective attention is compared with a regular
auto-encoder and previous works using a deep belief network (DBN).
| 0 | 0 | 0 | 0 | 1 | 0 |
Laser Wakefield Accelerators | The one-dimensional wakefield generation equations are solved for increasing
levels of non-linearity, to demonstrate how they contribute to the overall
behaviour of a non-linear wakefield in a plasma. The effect of laser guiding is
also studied as a way to increase the interaction length of a laser wakefield
accelerator.
| 0 | 1 | 0 | 0 | 0 | 0 |
Permutation complexity of images of Sturmian words by marked morphisms | We show that the permutation complexity of the image of a Sturmian word by a
binary marked morphism is $n+k$ for some constant $k$ and all lengths $n$
sufficiently large.
| 1 | 0 | 0 | 0 | 0 | 0 |
Three principles of data science: predictability, computability, and stability (PCS) | We propose the predictability, computability, and stability (PCS) framework
to extract reproducible knowledge from data that can guide scientific
hypothesis generation and experimental design. The PCS framework builds on key
ideas in machine learning, using predictability as a reality check and
evaluating computational considerations in data collection, data storage, and
algorithm design. It augments PC with an overarching stability principle, which
largely expands traditional statistical uncertainty considerations. In
particular, stability assesses how results vary with respect to choices (or
perturbations) made across the data science life cycle, including problem
formulation, pre-processing, modeling (data and algorithm perturbations), and
exploratory data analysis (EDA) before and after modeling.
Furthermore, we develop PCS inference to investigate the stability of data
results and identify when models are consistent with relatively simple
phenomena. We compare PCS inference with existing methods, such as selective
inference, in high-dimensional sparse linear model simulations to demonstrate
that our methods consistently outperform others in terms of ROC curves over a
wide range of simulation settings. Finally, we propose a PCS documentation
based on Rmarkdown, iPython, or Jupyter Notebook, with publicly available,
reproducible codes and narratives to back up human choices made throughout an
analysis. The PCS workflow and documentation are demonstrated in a genomics
case study available on Zenodo.
| 1 | 0 | 0 | 1 | 0 | 0 |
The Elasticity of Puiseux Monoids | Let $M$ be an atomic monoid and let $x$ be a non-unit element of $M$. The
elasticity of $x$, denoted by $\rho(x)$, is the ratio of its largest
factorization length to its shortest factorization length, and it measures how
far is $x$ from having a unique factorization. The elasticity $\rho(M)$ of $M$
is the supremum of the elasticities of all non-unit elements of $M$. The monoid
$M$ has accepted elasticity if $\rho(M) = \rho(m)$ for some $m \in M$. In this
paper, we study the elasticity of Puiseux monoids (i.e., additive submonoids of
$\mathbb{Q}_{\ge 0}$). First, we characterize the Puiseux monoids $M$ having
finite elasticity and find a formula for $\rho(M)$. Then we classify the
Puiseux monoids having accepted elasticity in terms of their sets of atoms.
When $M$ is a primary Puiseux monoid, we describe the topology of the set of
elasticities of $M$, including a characterization of when $M$ is a bounded
factorization monoid. Lastly, we give an example of a Puiseux monoid that is
bifurcus (that is, every nonzero element has a factorization of length at most
$2$).
| 0 | 0 | 1 | 0 | 0 | 0 |
Barcode Embeddings for Metric Graphs | Stable topological invariants are a cornerstone of persistence theory and
applied topology, but their discriminative properties are often
poorly-understood. In this paper we study a rich homology-based invariant first
defined by Dey, Shi, and Wang, which we think of as embedding a metric graph in
the barcode space. We prove that this invariant is locally injective on the
space of metric graphs and globally injective on a GH-dense subset. Moreover,
we define a new topology on the space of metric graphs, which we call the
fibered topology, for which the barcode transform is injective on a generic
(open and dense) subset.
| 0 | 0 | 1 | 0 | 0 | 0 |
On the extremal extensions of a non-negative Jacobi operator | We consider minimal non-negative Jacobi operator with $p\times p-$matrix
entries. Using the technique of boundary triplets and the corresponding Weyl
functions, we describe the Friedrichs and Krein extensions of the minimal
Jacobi operator. Moreover, we parameterize the set of all non-negative
self-adjoint extensions in terms of boundary conditions.
| 0 | 0 | 1 | 0 | 0 | 0 |
Note on the geodesic Monte Carlo | Geodesic Monte Carlo (gMC) is a powerful algorithm for Bayesian inference on
non-Euclidean manifolds. The original gMC algorithm was cleverly derived in
terms of its progenitor, the Riemannian manifold Hamiltonian Monte Carlo
(RMHMC). Here, it is shown that alternative and theoretically simpler
derivations are available in which the original algorithm is a special case of
two general classes of algorithms characterized by non-trivial mass matrices.
The proposed derivations work entirely in embedding coordinates and thus
clarify the original algorithm as applied to manifolds embedded in Euclidean
space.
| 0 | 0 | 0 | 1 | 0 | 0 |
Google Scholar and the gray literature: A reply to Bonato's review | Recently, a review concluded that Google Scholar (GS) is not a suitable
source of information "for identifying recent conference papers or other gray
literature publications". The goal of this letter is to demonstrate that GS can
be an effective tool to search and find gray literature, as long as appropriate
search strategies are used. To do this, we took as examples the same two case
studies used by the original review, describing first how GS processes
original's search strategies, then proposing alternative search strategies, and
finally generalizing each case study to compose a general search procedure
aimed at finding gray literature in Google Scholar for two wide selected case
studies: a) all contributions belonging to a congress (the ASCO Annual
Meeting); and b) indexed guidelines as well as gray literature within medical
institutions (National Institutes of Health) and governmental agencies (U.S.
Department of Health & Human Services). The results confirm that original
search strategies were undertrained offering misleading results and erroneous
conclusions. Google Scholar lacks many of the advanced search features
available in other bibliographic databases (such as Pubmed), however, it is one
thing to have a friendly search experience, and quite another to find gray
literature. We finally conclude that Google Scholar is a powerful tool for
searching gray literature, as long as the users are familiar with all the
possibilities it offers as a search engine. Poorly formulated searches will
undoubtedly return misleading results.
| 1 | 0 | 0 | 0 | 0 | 0 |
A Next-Best-Smell Approach for Remote Gas Detection with a Mobile Robot | The problem of gas detection is relevant to many real-world applications,
such as leak detection in industrial settings and landfill monitoring. Using
mobile robots for gas detection has several advantages and can reduce danger
for humans. In our work, we address the problem of planning a path for a mobile
robotic platform equipped with a remote gas sensor, which minimizes the time to
detect all gas sources in a given environment. We cast this problem as a
coverage planning problem by defining a basic sensing operation -- a scan with
the remote gas sensor -- as the field of "view" of the sensor. Given the
computing effort required by previously proposed offline approaches, in this
paper we suggest a online coverage algorithm, called Next-Best-Smell, adapted
from the Next-Best-View class of exploration algorithms. Our algorithm
evaluates candidate locations with a global utility function, which combines
utility values for travel distance, information gain, and sensing time, using
Multi-Criteria Decision Making. In our experiments, conducted both in
simulation and with a real robot, we found the performance of the
Next-Best-Smell approach to be comparable with that of the state-of-the-art
offline algorithm, at much lower computational cost.
| 1 | 0 | 0 | 0 | 0 | 0 |
KZ-Calogero correspondence revisited | We discuss the correspondence between the Knizhnik-Zamolodchikov equations
associated with $GL(N)$ and the $n$-particle quantum Calogero model in the case
when $n$ is not necessarily equal to $N$. This can be viewed as a natural
"quantization" of the quantum-classical correspondence between quantum Gaudin
and classical Calogero models.
| 0 | 1 | 1 | 0 | 0 | 0 |
Spectral up- and downshifting of Akhmediev breathers under wind forcing | We experimentally and numerically investigate the effect of wind forcing on
the spectral dynamics of Akhmediev breathers, a wave-type known to model the
modulation instability. We develop the wind model to the same order in
steepness as the higher order modifcation of the nonlinear Schroedinger
equation, also referred to as the Dysthe equation. This results in an
asymmetric wind term in the higher order, in addition to the leading order wind
forcing term. The derived model is in good agreement with laboratory
experiments within the range of the facility's length. We show that the leading
order forcing term amplifies all frequencies equally and therefore induces only
a broadening of the spectrum while the asymmetric higher order term in the
model enhances higher frequencies more than lower ones. Thus, the latter term
induces a permanent upshift of the spectral mean. On the other hand, in
contrast to the direct effect of wind forcing, wind can indirectly lead to
frequency downshifts, due to dissipative effects such as wave breaking, or
through amplification of the intrinsic spectral asymmetry of the Dysthe
equation. Furthermore, the definitions of the up- and downshift in terms of
peak- and mean frequencies, that are critical to relate our work to previous
results, are highlighted and discussed.
| 0 | 1 | 0 | 0 | 0 | 0 |
High Rate LDPC Codes from Difference Covering Arrays | This paper presents a combinatorial construction of low-density parity-check
(LDPC) codes from difference covering arrays. While the original construction
by Gallagher was by randomly allocating bits in a sparse parity-check matrix,
over the past 20 years researchers have used a variety of more structured
approaches to construct these codes, with the more recent constructions of
well-structured LDPC coming from balanced incomplete block designs (BIBDs) and
from Latin squares over finite fields. However these constructions have
suffered from the limited orders for which these designs exist. Here we present
a construction of LDPC codes of length $4n^2 - 2n$ for all $n$ using the cyclic
group of order $2n$. These codes achieve high information rate (greater than
0.8) for $n \geq 8$, have girth at least 6 and have minimum distance 6 for $n$
odd.
| 1 | 0 | 1 | 0 | 0 | 0 |
Bootstrapping spectral statistics in high dimensions | Statistics derived from the eigenvalues of sample covariance matrices are
called spectral statistics, and they play a central role in multivariate
testing. Although bootstrap methods are an established approach to
approximating the laws of spectral statistics in low-dimensional problems,
these methods are relatively unexplored in the high-dimensional setting. The
aim of this paper is to focus on linear spectral statistics as a class of
prototypes for developing a new bootstrap in high-dimensions --- and we refer
to this method as the Spectral Bootstrap. In essence, the method originates
from the parametric bootstrap, and is motivated by the notion that, in high
dimensions, it is difficult to obtain a non-parametric approximation to the
full data-generating distribution. From a practical standpoint, the method is
easy to use, and allows the user to circumvent the difficulties of complex
asymptotic formulas for linear spectral statistics. In addition to proving the
consistency of the proposed method, we provide encouraging empirical results in
a variety of settings. Lastly, and perhaps most interestingly, we show through
simulations that the method can be applied successfully to statistics outside
the class of linear spectral statistics, such as the largest sample eigenvalue
and others.
| 0 | 0 | 0 | 1 | 0 | 0 |
On parabolic subgroups of Artin-Tits groups of spherical type | We show that, in an Artin-Tits group of spherical type, the intersection of
two parabolic subgroups is a parabolic subgroup. Moreover, we show that the set
of parabolic subgroups forms a lattice with respect to inclusion. This extends
to all Artin-Tits groups of spherical type a result that was previously known
for braid groups.
To obtain the above results, we show that every element in an Artin-Tits
group of spherical type admits a unique minimal parabolic subgroup containing
it. Also, the subgroup associated to an element coincides with the subgroup
associated to any of its powers or roots. As a consequence, if an element
belongs to a parabolic subgroup, all its roots belong to the same parabolic
subgroup.
We define the simplicial complex of irreducible parabolic subgroups, and we
propose it as the analogue, in Artin-Tits groups of spherical type, of the
celebrated complex of curves which is an important tool in braid groups, and
more generally in mapping class groups. We conjecture that the complex of
irreducible parabolic subgroups is $\delta$-hyperbolic.
| 0 | 0 | 1 | 0 | 0 | 0 |
A statistical test for the Zipf's law by deviations from the Heaps' law | We explore a probabilistic model of an artistic text: words of the text are
chosen independently of each other in accordance with a discrete probability
distribution on an infinite dictionary. The words are enumerated 1, 2,
$\ldots$, and the probability of appearing the $i$'th word is asymptotically a
power function. Bahadur proved that in this case the number of different words
depends on the length of the text is asymptotically a power function, too. On
the other hand, in the applied statistics community, there exist statements
supported by empirical observations, the Zipf's and the Heaps' laws. We
highlight the links between Bahadur results and Zipf's/Heaps' laws, and
introduce and analyse a corresponding statistical test.
| 0 | 0 | 1 | 1 | 0 | 0 |
Weakly Supervised Classification in High Energy Physics | As machine learning algorithms become increasingly sophisticated to exploit
subtle features of the data, they often become more dependent on simulations.
This paper presents a new approach called weakly supervised classification in
which class proportions are the only input into the machine learning algorithm.
Using one of the most challenging binary classification tasks in high energy
physics - quark versus gluon tagging - we show that weakly supervised
classification can match the performance of fully supervised algorithms.
Furthermore, by design, the new algorithm is insensitive to any mis-modeling of
discriminating features in the data by the simulation. Weakly supervised
classification is a general procedure that can be applied to a wide variety of
learning problems to boost performance and robustness when detailed simulations
are not reliable or not available.
| 0 | 1 | 0 | 1 | 0 | 0 |
Profit Driven Decision Trees for Churn Prediction | Customer retention campaigns increasingly rely on predictive models to detect
potential churners in a vast customer base. From the perspective of machine
learning, the task of predicting customer churn can be presented as a binary
classification problem. Using data on historic behavior, classification
algorithms are built with the purpose of accurately predicting the probability
of a customer defecting. The predictive churn models are then commonly selected
based on accuracy related performance measures such as the area under the ROC
curve (AUC). However, these models are often not well aligned with the core
business requirement of profit maximization, in the sense that, the models fail
to take into account not only misclassification costs, but also the benefits
originating from a correct classification. Therefore, the aim is to construct
churn prediction models that are profitable and preferably interpretable too.
The recently developed expected maximum profit measure for customer churn
(EMPC) has been proposed in order to select the most profitable churn model. We
present a new classifier that integrates the EMPC metric directly into the
model construction. Our technique, called ProfTree, uses an evolutionary
algorithm for learning profit driven decision trees. In a benchmark study with
real-life data sets from various telecommunication service providers, we show
that ProfTree achieves significant profit improvements compared to classic
accuracy driven tree-based methods.
| 1 | 0 | 0 | 1 | 0 | 0 |
Carrier loss mechanisms in textured crystalline Si-based solar cells | A quite general device analysis method that allows the direct evaluation of
optical and recombination losses in crystalline silicon (c-Si)-based solar
cells has been developed. By applying this technique, the optical and physical
limiting factors of the state-of-the-art solar cells with ~20% efficiencies
have been revealed. In the established method, the carrier loss mechanisms are
characterized from the external quantum efficiency (EQE) analysis with very low
computational cost. In particular, the EQE analyses of textured c-Si solar
cells are implemented by employing the experimental reflectance spectra
obtained directly from the actual devices while using flat optical models
without any fitting parameters. We find that the developed method provides
almost perfect fitting to EQE spectra reported for various textured c-Si solar
cells, including c-Si heterojunction solar cells, a dopant-free c-Si solar cell
with a MoOx layer, and an n-type passivated emitter with rear locally diffused
(PERL) solar cell. The modeling of the recombination loss further allows the
extraction of the minority carrier diffusion length and surface recombination
velocity from the EQE analysis. Based on the EQE analysis results, the carrier
loss mechanisms in different types of c-Si solar cells are discussed.
| 0 | 1 | 0 | 0 | 0 | 0 |
The generating function for the Airy point process and a system of coupled Painlevé II equations | For a wide class of Hermitian random matrices, the limit distribution of the
eigenvalues close to the largest one is governed by the Airy point process. In
such ensembles, the limit distribution of the k-th largest eigenvalue is given
in terms of the Airy kernel Fredholm determinant or in terms of Tracy-Widom
formulas involving solutions of the Painlevé II equation. Limit distributions
for quantities involving two or more near-extreme eigenvalues, such as the gap
between the k-th and the \ell-th largest eigenvalue or the sum of the k largest
eigenvalues, can be expressed in terms of Fredholm determinants of an Airy
kernel with several discontinuities. We establish simple Tracy-Widom type
expressions for these Fredholm determinants, which involve solutions to systems
of coupled Painlevé II equations, and we investigate the asymptotic behavior
of these solutions.
| 0 | 0 | 1 | 0 | 0 | 0 |
Gradient-based Training of Slow Feature Analysis by Differentiable Approximate Whitening | This paper proposes Power Slow Feature Analysis, a gradient-based method to
extract temporally-slow features from a high-dimensional input stream that
varies on a faster time-scale, and a variant of Slow Feature Analysis (SFA).
While displaying performance comparable to hierarchical extensions to the SFA
algorithm, such as Hierarchical Slow Feature Analysis, for a small number of
output-features, our algorithm allows end-to-end training of arbitrary
differentiable approximators (e.g., deep neural networks). We provide
experimental evidence that PowerSFA is able to extract meaningful and
informative low-dimensional features in the case of a) synthetic
low-dimensional data, b) visual data, and also for c) a general dataset for
which symmetric non-temporal relations between points can be defined.
| 0 | 0 | 0 | 1 | 0 | 0 |
Quantum Critical Behavior in the Asymptotic Limit of High Disorder: Entropy Stabilized NiCoCr0.8 Alloys | The behavior of matter near a quantum critical point (QCP) is one of the most
exciting and challenging areas of physics research. Emergent phenomena such as
high-temperature superconductivity are linked to the proximity to a QCP.
Although significant progress has been made in understanding quantum critical
behavior in some low dimensional magnetic insulators, the situation in metallic
systems is much less clear. Here we demonstrate that NiCoCrx single crystal
alloys are remarkable model systems for investigating QCP physics in a metallic
environment. For NiCoCrx alloys with x = 0.8, the critical exponents associated
with a ferromagnetic quantum critical point (FQCP) are experimentally
determined from low temperature magnetization and heat capacity measurements.
For the first time, all of the five critical exponents ( gamma-subT =1/2 ,
beta-subT = 1, delta = 3/2, nuz-subm = 2, alpha-bar-subT = 0) are in remarkable
agreement with predictions of Belitz-Kirkpatrick-Vojta (BKV) theory in the
asymptotic limit of high disorder. Using these critical exponents, excellent
scaling of the magnetization data is demonstrated with no adjustable
parameters. We also find a divergence of the magnetic Gruneisen parameter,
consistent with a FQCP. This work therefore demonstrates that entropy
stabilized concentrated solid solutions represent a unique platform to study
quantum critical behavior in a highly tunable class of materials.
| 0 | 1 | 0 | 0 | 0 | 0 |
Multilayer Network Modeling of Integrated Biological Systems | Biological systems, from a cell to the human brain, are inherently complex. A
powerful representation of such systems, described by an intricate web of
relationships across multiple scales, is provided by complex networks.
Recently, several studies are highlighting how simple networks -- obtained by
aggregating or neglecting temporal or categorical description of biological
data -- are not able to account for the richness of information characterizing
biological systems. More complex models, namely multilayer networks, are needed
to account for interdependencies, often varying across time, of biological
interacting units within a cell, a tissue or parts of an organism.
| 0 | 0 | 0 | 0 | 1 | 0 |
Fast, Autonomous Flight in GPS-Denied and Cluttered Environments | One of the most challenging tasks for a flying robot is to autonomously
navigate between target locations quickly and reliably while avoiding obstacles
in its path, and with little to no a-priori knowledge of the operating
environment. This challenge is addressed in the present paper. We describe the
system design and software architecture of our proposed solution, and showcase
how all the distinct components can be integrated to enable smooth robot
operation. We provide critical insight on hardware and software component
selection and development, and present results from extensive experimental
testing in real-world warehouse environments. Experimental testing reveals that
our proposed solution can deliver fast and robust aerial robot autonomous
navigation in cluttered, GPS-denied environments.
| 1 | 0 | 0 | 0 | 0 | 0 |
Accumulation Bit-Width Scaling For Ultra-Low Precision Training Of Deep Networks | Efforts to reduce the numerical precision of computations in deep learning
training have yielded systems that aggressively quantize weights and
activations, yet employ wide high-precision accumulators for partial sums in
inner-product operations to preserve the quality of convergence. The absence of
any framework to analyze the precision requirements of partial sum
accumulations results in conservative design choices. This imposes an
upper-bound on the reduction of complexity of multiply-accumulate units. We
present a statistical approach to analyze the impact of reduced accumulation
precision on deep learning training. Observing that a bad choice for
accumulation precision results in loss of information that manifests itself as
a reduction in variance in an ensemble of partial sums, we derive a set of
equations that relate this variance to the length of accumulation and the
minimum number of bits needed for accumulation. We apply our analysis to three
benchmark networks: CIFAR-10 ResNet 32, ImageNet ResNet 18 and ImageNet
AlexNet. In each case, with accumulation precision set in accordance with our
proposed equations, the networks successfully converge to the single precision
floating-point baseline. We also show that reducing accumulation precision
further degrades the quality of the trained network, proving that our equations
produce tight bounds. Overall this analysis enables precise tailoring of
computation hardware to the application, yielding area- and power-optimal
systems.
| 1 | 0 | 0 | 1 | 0 | 0 |
The Berkovich realization for rigid analytic motives | We prove that the functor associating to a rigid analytic variety the
singular complex of the underlying Berkovich topological space is motivic, and
defines the maximal Artin quotient of a motive. We use this to generalize
Berkovich's results on the weight-zero part of the étale cohomology of a
variety defined over a non-archimedean valued field.
| 0 | 0 | 1 | 0 | 0 | 0 |
Rates of estimation for determinantal point processes | Determinantal point processes (DPPs) have wide-ranging applications in
machine learning, where they are used to enforce the notion of diversity in
subset selection problems. Many estimators have been proposed, but surprisingly
the basic properties of the maximum likelihood estimator (MLE) have received
little attention. In this paper, we study the local geometry of the expected
log-likelihood function to prove several rates of convergence for the MLE. We
also give a complete characterization of the case where the MLE converges at a
parametric rate. Even in the latter case, we also exhibit a potential curse of
dimensionality where the asymptotic variance of the MLE is exponentially large
in the dimension of the problem.
| 0 | 0 | 1 | 1 | 0 | 0 |
Negative refraction in Weyl semimetals | We theoretically propose that Weyl semimetals may exhibit negative refraction
at some frequencies close to the plasmon frequency, allowing transverse
magnetic (TM) electromagnetic waves with frequencies smaller than the plasmon
frequency to propagate in the Weyl semimetals. The idea is justified by the
calculation of reflection spectra, in which \textit{negative} refractive index
at such frequencies gives physically correct spectra. In this case, a TM
electromagnetic wave incident to the surface of the Weyl semimetal will be bent
with a negative angle of refraction. We argue that the negative refractive
index at the specified frequencies of the electromagnetic wave is required to
conserve the energy of the wave, in which the incident energy should propagate
away from the point of incidence.
| 0 | 1 | 0 | 0 | 0 | 0 |
Towards Object Life Cycle-Based Variant Generation of Business Process Models | Variability management of process models is a major challenge for
Process-Aware Information Systems. Process model variants can be attributed to
any of the following reasons: new technologies, governmental rules,
organizational context or adoption of new standards. Current approaches to
manage variants of process models address issues such as reducing the huge
effort of modeling from scratch, preventing redundancy, and controlling
inconsistency in process models. Although the effort to manage process model
variants has been exerted, there are still limitations. Furthermore, existing
approaches do not focus on variants that come from change in organizational
perspective of process models. Organizational-driven variant management is an
important area that still needs more study that we focus on in this paper.
Object Life Cycle (OLC) is an important aspect that may change from an
organization to another. This paper introduces an approach inspired by real
life scenario to generate consistent process model variants that come from
adaptations in the OLC.
| 1 | 0 | 0 | 0 | 0 | 0 |
Detection of sub-MeV Dark Matter with Three-Dimensional Dirac Materials | We propose the use of three-dimensional Dirac materials as targets for direct
detection of sub-MeV dark matter. Dirac materials are characterized by a linear
dispersion for low-energy electronic excitations, with a small band gap of
O(meV) if lattice symmetries are broken. Dark matter at the keV scale carrying
kinetic energy as small as a few meV can scatter and excite an electron across
the gap. Alternatively, bosonic dark matter as light as a few meV can be
absorbed by the electrons in the target. We develop the formalism for dark
matter scattering and absorption in Dirac materials and calculate the
experimental reach of these target materials. We find that Dirac materials can
play a crucial role in detecting dark matter in the keV to MeV mass range that
scatters with electrons via a kinetically mixed dark photon, as the dark photon
does not develop an in-medium effective mass. The same target materials provide
excellent sensitivity to absorption of light bosonic dark matter in the meV to
hundreds of meV mass range, superior to all other existing proposals when the
dark matter is a kinetically mixed dark photon.
| 0 | 1 | 0 | 0 | 0 | 0 |
Formation of Local Resonance Band Gaps in Finite Acoustic Metamaterials: A Closed-form Transfer Function Model | The objective of this paper is to use transfer functions to comprehend the
formation of band gaps in locally resonant acoustic metamaterials. Identifying
a recursive approach for any number of serially arranged locally resonant mass
in mass cells, a closed form expression for the transfer function is derived.
Analysis of the end-to-end transfer function helps identify the fundamental
mechanism for the band gap formation in a finite metamaterial. This mechanism
includes (a) repeated complex conjugate zeros located at the natural frequency
of the individual local resonators, (b) the presence of two poles which flank
the band gap, and (c) the absence of poles in the band-gap. Analysis of the
finite cell dynamics are compared to the Bloch-wave analysis of infinitely long
metamaterials to confirm the theoretical limits of the band gap estimated by
the transfer function modeling. The analysis also explains how the band gap
evolves as the number of cells in the metamaterial chain increases and
highlights how the response varies depending on the chosen sensing location
along the length of the metamaterial. The proposed transfer function approach
to compute and evaluate band gaps in locally resonant structures provides a
framework for the exploitation of control techniques to modify and tune band
gaps in finite metamaterial realizations.
| 0 | 1 | 1 | 0 | 0 | 0 |
Online Multi-Label Classification: A Label Compression Method | Many modern applications deal with multi-label data, such as functional
categorizations of genes, image labeling and text categorization.
Classification of such data with a large number of labels and latent
dependencies among them is a challenging task, and it becomes even more
challenging when the data is received online and in chunks. Many of the current
multi-label classification methods require a lot of time and memory, which make
them infeasible for practical real-world applications. In this paper, we
propose a fast linear label space dimension reduction method that transforms
the labels into a reduced encoded space and trains models on the obtained
pseudo labels. Additionally, it provides an analytical method to update the
decoding matrix which maps the labels into the original space and is used
during the test phase. Experimental results show the effectiveness of this
approach in terms of running times and the prediction performance over
different measures.
| 0 | 0 | 0 | 1 | 0 | 0 |
Mutual Alignment Transfer Learning | Training robots for operation in the real world is a complex, time consuming
and potentially expensive task. Despite significant success of reinforcement
learning in games and simulations, research in real robot applications has not
been able to match similar progress. While sample complexity can be reduced by
training policies in simulation, such policies can perform sub-optimally on the
real platform given imperfect calibration of model dynamics. We present an
approach -- supplemental to fine tuning on the real robot -- to further benefit
from parallel access to a simulator during training and reduce sample
requirements on the real robot. The developed approach harnesses auxiliary
rewards to guide the exploration for the real world agent based on the
proficiency of the agent in simulation and vice versa. In this context, we
demonstrate empirically that the reciprocal alignment for both agents provides
further benefit as the agent in simulation can adjust to optimize its behaviour
for states commonly visited by the real-world agent.
| 1 | 0 | 0 | 0 | 0 | 0 |
Tangent: Automatic differentiation using source-code transformation for dynamically typed array programming | The need to efficiently calculate first- and higher-order derivatives of
increasingly complex models expressed in Python has stressed or exceeded the
capabilities of available tools. In this work, we explore techniques from the
field of automatic differentiation (AD) that can give researchers expressive
power, performance and strong usability. These include source-code
transformation (SCT), flexible gradient surgery, efficient in-place array
operations, higher-order derivatives as well as mixing of forward and reverse
mode AD. We implement and demonstrate these ideas in the Tangent software
library for Python, the first AD framework for a dynamic language that uses
SCT.
| 1 | 0 | 0 | 0 | 0 | 0 |
Scalar and Tensor Parameters for Importing Tensor Index Notation including Einstein Summation Notation | In this paper, we propose a method for importing tensor index notation,
including Einstein summation notation, into functional programming. This method
involves introducing two types of parameters, i.e, scalar and tensor
parameters, and simplified tensor index rules that do not handle expressions
that are valid only for the Cartesian coordinate system, in which the index can
move up and down freely. An example of such an expression is "c = A_i B_i". As
an ordinary function, when a tensor parameter obtains a tensor as an argument,
the function treats the tensor argument as a whole. In contrast, when a scalar
parameter obtains a tensor as an argument, the function is applied to each
component of the tensor. In this paper, we show that introducing these two
types of parameters and our simplified index rules enables us to apply
arbitrary user-defined functions to tensor arguments using index notation
including Einstein summation notation without requiring an additional
description to enable each function to handle tensors.
| 1 | 0 | 0 | 0 | 0 | 0 |
Evidence for coherent spicule oscillations by correcting Hinode/SOT Ca II H in the southeast limb of the Sun | Wave theories of heating the chromosphere, corona, and solar wind due to
photospheric fluctuations are strengthened by the existence of observed wave
coherency up to the transition region (TR). The coherency of solar spicules'
intensity oscillations was explored using the Solar Optical Telescope (SOT) on
the Hinode spacecraft with a height increase above the solar limb in active
region (AR). We used time sequences near the southeast region from the
Hinode/SOT in Ca II H line obtained on April 3, 2015 and applied the
de-convolution procedure to the spicule in order to illustrate how effectively
our restoration method works on fine structures such as spicules. Moreover, the
intensity oscillations at different heights above the solar limb were analysed
through wavelet transforms. Afterwards, the phase difference was measured among
oscillations at two certain heights in search of evidence for coherent
oscillations. The results of wavelet transformations revealed dominant period
peaks in 2, 4, 5.5, and 6.5 min at four separate heights. The dominant
frequencies for coherency level higher than 75 percent was found to be around
5.5 and 8.5 mHz. Mean phase speeds of 155-360 km s-1 were measured. We found
that the mean phase speeds increased with height. The results suggest that the
energy flux carried by coherent waves into the corona and heliosphere may be
several times larger than previous estimates that were based solely on constant
velocities. We provide compelling evidence for the existence of upwardly
propagating coherent waves.
| 0 | 1 | 0 | 0 | 0 | 0 |
Dynamics and transverse relaxation of an unconventional spin-rotation mode in a two-dimensional strongly magnetized electron gas | An unconventional spin-rotation mode emerging in a quantum Hall ferromagnet
due to excitation by a laser pulse is studied. This state, macroscopically
representing a rotation of the entire electron spin-system to a certain angle,
microscopically is not equivalent to a coherent turn of all spins as a
single-whole and is presented in the form of a combination of eigen quantum
states corresponding to all possible S_z spin numbers. Motion of the
macroscopic quantum state is studied microscopically by solving a
non-stationary Schroedinger equation and by means of a kinetic approach where
damping of the spin-rotation mode is related to an elementary process -
transformation of a 'Goldstone spin exciton' to a 'spin-wave exciton'. The
system exhibits a spin stochastization mechanism (determined by spatial
fluctuations of the g-factor) providing the damping - the transverse spin
relaxation, but irrelevant to a decay of spin-wave excitons and thus not
providing the longitudinal relaxation - recovery of the S_z number to its
equilibrium value.
| 0 | 1 | 0 | 0 | 0 | 0 |
Opinion diversity and community formation in adaptive networks | It is interesting and of significant importance to investigate how network
structures co-evolve with opinions. The existing models of such co-evolution
typically lead to the final states where network nodes either reach a global
consensus or break into separated communities, each of which holding its own
community consensus. Such results, however, can hardly explain the richness of
real-life observations that opinions are always diversified with no global or
even community consensus, and people seldom, if not never, totally cut off
themselves from dissenters. In this article, we show that, a simple model
integrating consensus formation, link rewiring and opinion change allows
complex system dynamics to emerge, driving the system into a dynamic
equilibrium with co-existence of diversified opinions. Specifically, similar
opinion holders may form into communities yet with no strict community
consensus; and rather than being separated into disconnected communities,
different communities remain to be interconnected by non-trivial proportion of
inter-community links. More importantly, we show that the complex dynamics may
lead to different numbers of communities at steady state with a given tolerance
between different opinion holders. We construct a framework for theoretically
analyzing the co-evolution process. Theoretical analysis and extensive
simulation results reveal some useful insights into the complex co-evolution
process, including the formation of dynamic equilibrium, the phase transition
between different steady states with different numbers of communities, and the
dynamics between opinion distribution and network modularity, etc.
| 1 | 1 | 0 | 0 | 0 | 0 |
On the local and global comparison of generalized Bajraktarević means | Given two continuous functions $f,g:I\to\mathbb{R}$ such that $g$ is positive
and $f/g$ is strictly monotone, a measurable space $(T,A)$, a measurable family
of $d$-variable means $m: I^d\times T\to I$, and a probability measure $\mu$ on
the measurable sets $A$, the $d$-variable mean $M_{f,g,m;\mu}:I^d\to I$ is
defined by $$
M_{f,g,m;\mu}(\pmb{x})
:=\left(\frac{f}{g}\right)^{-1}\left(
\frac{\int_T f\big(m(x_1,\dots,x_d,t)\big) d\mu(t)}
{\int_T g\big(m(x_1,\dots,x_d,t)\big) d\mu(t)}\right)
\qquad(\pmb{x}=(x_1,\dots,x_d)\in I^d). $$ The aim of this paper is to study
the local and global comparison problem of these means, i.e., to find
conditions for the generating functions $(f,g)$ and $(h,k)$, for the families
of means $m$ and $n$, and for the measures $\mu,\nu$ such that the comparison
inequality $$
M_{f,g,m;\mu}(\pmb{x})\leq M_{h,k,n;\nu}(\pmb{x}) \qquad(\pmb{x}\in I^d) $$
be satisfied.
| 0 | 0 | 1 | 0 | 0 | 0 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.