text
stringlengths 6
128k
|
|---|
Topological properties of "scale-free" networks are investigated by
determining their spectral dimensions $d_S$, which reflect a diffusion process
in the corresponding graphs. Data bases for citation networks and metabolic
networks together with simulation results from the growing network model
\cite{barab} are probed. For completeness and comparisons lattice, random,
small-world models are also investigated. We find that $d_S$ is around 3 for
citation and metabolic networks, which is significantly different from the
growing network model, for which $d_S$ is approximately 7.5. This signals a
substantial difference in network topology despite the observed similarities in
vertex order distributions. In addition, the diffusion analysis indicates that
whereas the citation networks are tree-like in structure, the metabolic
networks contain many loops.
|
As Artificial Intelligence (AI) technology becomes more and more prevalent,
it becomes increasingly important to explore how we as humans interact with AI.
The Human-AI Interaction (HAI) sub-field has emerged from the Human-Computer
Interaction (HCI) field and aims to examine this very notion. Many interaction
patterns have been implemented without fully understanding the changes in
required cognition as well as the cognitive science implications of using these
alternative interfaces that aim to be more human-like in nature. Prior research
suggests that theory of mind representations are crucial to successful and
effortless communication, however very little is understood when it comes to
how theory of mind representations are established when interacting with AI.
|
We report trigonometric parallaxes for the high-mass star forming regions
G35.20-0.74 and G35.20-1.74, corresponding to distances of 2.19 (+0.24 -0.20)
kpc and 3.27 (+0.56 -0.42) kpc, respectively. The distances to both sources are
close to their near kinematic distances and place them in the
Carina-Sagittarius spiral arm. Combining the distances and proper motions with
observed radial velocities gives the locations and full space motions of the
star forming regions. Assuming a standard model of the Galaxy, G35.20-0.74 and
G35.20-1.74 have peculiar motions of ~13 km/s and ~16 km/s counter to Galactic
rotation and ~9 km/s toward the North Galactic Pole.
|
This article is devoted to the study of tail index estimation based on i.i.d.
multivariate observations, drawn from a standard heavy-tailed distribution,
i.e. of which 1-d Pareto-like marginals share the same tail index. A
multivariate Central Limit Theorem for a random vector, whose components
correspond to (possibly dependent) Hill estimators of the common shape index
alpha, is established under mild conditions. Motivated by the statistical
analysis of extremal spatial data in particular, we introduce the concept of
(standard) heavy-tailed random field of tail index alpha and show how this
limit result can be used in order to build an estimator of alpha with small
asymptotic mean squared error, through a proper convex linear combination of
the coordinates. Beyond asymptotic results, simulation experiments illustrating
the relevance of the approach promoted are also presented.
|
We employ the G-structure formalism to study supersymmetric solutions of
minimal and SU(2) gauged supergravities in seven dimensions admitting Killing
spinors with associated timelike Killing vector. The most general such Killing
spinor defines an SU(3) structure. We deduce necessary and sufficient
conditions for the existence of a timelike Killing spinor on the bosonic fields
of the theories, and find that such configurations generically preserve one out
of sixteen supersymmetries. Using our general supersymmetric ansatz we obtain
numerous new solutions, including squashed or deformed AdS solutions of the
gauged theory, and a large class of Godel-like solutions with closed timelike
curves.
|
We extend the dynamical systems analysis of Scalar-Fluid interacting dark
energy models performed in C. G. Boehmer et al, Phys. Rev. D 91, 123002 (2015),
by considering scalar field potentials beyond the exponential type. The
properties and stability of critical points are examined using a combination of
linear analysis, computational methods and advanced mathematical techniques,
such as centre manifold theory. We show that the interesting results obtained
with an exponential potential can generally be recovered also for more
complicated scalar field potentials. In particular, employing power-law and
hyperbolic potentials as examples, we find late time accelerated attractors,
transitions from dark matter to dark energy domination with specific
distinguishing features, and accelerated scaling solutions capable of solving
the cosmic coincidence problem.
|
The latest generation of transistors are nanoscale devices whose performance
and reliability are limited by thermal noise in low-power applications.
Therefore developing efficient methods to compute the voltage and current
fluctuations in such non-linear electronic circuits is essential. Traditional
approaches commonly rely on adding Gaussian white noise to the macroscopic
dynamical circuit laws, but do not capture rare fluctuations and lead to
thermodynamic inconsistencies. A correct and thermodynamically consistent
approach can be achieved by describing single-electron transfers as Poisson
jump processes accounting for charging effects. But such descriptions can be
computationally demanding. To address this issue, we consider the macroscopic
limit which corresponds to scaling up the physical dimensions of the transistor
and resulting in an increase of the number of electrons on the conductors. In
this limit, the thermal fluctuations satisfy a Large Deviations Principle which
we show is also remarkably precise in settings involving only a few tens of
electrons, by comparing our results with Gillespie simulations and spectral
methods. Traditional approaches are recovered by resorting to an ad hoc
diffusive approximation introducing inconsistencies. To illustrate these
findings, we consider a low-power CMOS inverter, or NOT gate, which is a basic
primitive in electronic design. Voltage (resp. current) fluctuations are
obtained analytically (semi-analytically) and reveal interesting features.
|
We present centimeter and millimeter observations of gas and dust around IRAS
21391+5802, an intermediate-mass source embedded in the core of IC 1396N.
Continuum observations from 3.6 cm to 1.2 mm are used to study the embedded
objects and overall distribution of the dust, while molecular line observations
of CO, CS, and CH3OH are used to probe the structure and chemistry of the
outflows in the region. The continuum emission at centimeter and millimeter
wavelengths has been resolved into three sources separated about 15 arcsec from
each other, and with one of them, BIMA 2, associated with IRAS 21391+5802. The
dust emission around this source shows a very extended envelope, which accounts
for most of the circumstellar mass of 5.1 Msun. This source is powering a
strong molecular outflow, elongated in the E--W direction, which presents a
complex structure and kinematics. While at high outflow velocities the outflow
is clearly bipolar, at low outflow velocities the blueshifted and redshifted
emission are highly overlapping, and the strongest emission shows a V-shaped
morphology. The outflow as traced by CS and CH3OH exhibits two well
differentiated and clumpy lobes, with two prominent northern blueshifted and
redshifted clumps. The curved shape of the clumps and the spectral shape at
these positions are consistent with shocked material. In addition, CS and CH3OH
are strongly enhanced toward these positions with respect to typical quiescent
material abundances in other star-forming regions.
|
We study number-phase uncertainty in a laser-driven, effectively four-level
atomic system under electromagnetically induced transparency (EIT) and coherent
population trapping (CPT). Uncertainty is described using (entropic) knowledge
of the two complementary variables, namely, number and phase, where knowledge
is defined as the relative entropy with respect to a uniform distribution. In
the regime where the coupling and probe lasers are approximately of equal
strength, and the atom exists in a CPT state, there is coherence between the
ground states, and correspondingly large phase knowledge and lower number
knowledge. The situation is the opposite in the case where coupling laser is
much stronger and the atom exists in an EIT state. We study these effects also
in the presence of a higher-order nonlinear absorption, which is seen to
produce a dephasing effect.
|
We introduce and initiate the study of a family of higher rank matricial
ranges, taking motivation from hybrid classical and quantum error correction
coding theory and its operator algebra framework. In particular, for a noisy
quantum channel, a hybrid quantum error correcting code exists if and only if a
distinguished special case of the joint higher rank matricial range of the
error operators of the channel is non-empty. We establish bounds on Hilbert
space dimension in terms of properties of a tuple of operators that guarantee a
matricial range is non-empty, and hence additionally guarantee the existence of
hybrid codes for a given quantum channel. We also discuss when hybrid codes can
have advantages over quantum codes and present a number of examples.
|
In this work we demonstrate that generative adversarial networks (GANs) can
be used to generate realistic pervasive changes in remote sensing imagery, even
in an unpaired training setting. We investigate some transformation quality
metrics based on deep embedding of the generated and real images which enable
visualization and understanding of the training dynamics of the GAN, and may
provide a useful measure in terms of quantifying how distinguishable the
generated images are from real images. We also identify some artifacts
introduced by the GAN in the generated images, which are likely to contribute
to the differences seen between the real and generated samples in the deep
embedding feature space even in cases where the real and generated samples
appear perceptually similar.
|
Quantum algorithms can be analyzed in a query model to compute Boolean
functions where input is given in a black box, but the aim is to compute
function value for arbitrary input using as few queries as possible. In this
paper we concentrate on quantum query algorithm designing tasks. The main aim
of research was to find new efficient algorithms and develop general algorithm
designing techniques. We present several exact quantum query algorithms for
certain problems that are better than classical counterparts. Next we introduce
algorithm transformation methods that allow significant enlarging of sets of
exactly computable functions. Finally, we propose algorithm constructing
methods applicable for algorithms with specific properties that allow
constructing algorithms for more complex functions preserving acceptable error
probability and number of queries.
|
We calculate various thermodynamic quantities of vortex liquids in a layered
superconductor by using the nonperturbative parquet approximation method, which
was previously used to study the effect of thermal fluctuations in
two-dimensional vortex systems. We find there is a first-order transition
between two vortex liquid phases which differ in the magnitude of their
correlation lengths. As the coupling between the layers increases,the
first-order transition line ends at a critical point. We discuss the possible
relation between this critical end-point and the disappearance of the
first-order transition which is observed in experiments on high temperature
superconductors at low magnetic fields.
|
Bayesian optimization offers a sample-efficient framework for navigating the
exploration-exploitation trade-off in the vast design space of biological
sequences. Whereas it is possible to optimize the various properties of
interest jointly using a multi-objective acquisition function, such as the
expected hypervolume improvement (EHVI), this approach does not account for
objectives with a hierarchical dependency structure. We consider a common use
case where some regions of the Pareto frontier are prioritized over others
according to a specified $\textit{partial ordering}$ in the objectives. For
instance, when designing antibodies, we would like to maximize the binding
affinity to a target antigen only if it can be expressed in live cell culture
-- modeling the experimental dependency in which affinity can only be measured
for antibodies that can be expressed and thus produced in viable quantities. In
general, we may want to confer a partial ordering to the properties such that
each property is optimized conditioned on its parent properties satisfying some
feasibility condition. To this end, we present PropertyDAG, a framework that
operates on top of the traditional multi-objective BO to impose this desired
ordering on the objectives, e.g. expression $\rightarrow$ affinity. We
demonstrate its performance over multiple simulated active learning iterations
on a penicillin production task, toy numerical problem, and a real-world
antibody design task.
|
We present the preliminary results of a deep (167 ks) ROSAT HRI observation
of the cD galaxy NGC1399 in the Fornax cluster. We find, in agreement with
previous observations, an extended (41 Kpc adopting a distance of 19 Mpc)
gaseous halo with a luminosity of L_X=(4.41\pm 0.04)x10^{41} erg/s. The 5
arcsec resolution of the data allows us to detect a very complex and asymmetric
structure of the halo with respect to the optical galaxy. Moreover the analysis
of the radial structure reveals the presence of a multi-component profile not
consistent with a simple King model over the whole 40 Kpc. We do not detect the
presence of a central source and pose an upper limit to the luminosity of a
possible active nucleus. Due to the length of the observation, comparable to
that of a deep survey, we detect a large number of sources within the HRI FOV,
in slight excess with respect to the estimates based on previous surveys. We
study the flux distribution of the sources, their temporal behaviour and their
spatial distribution with respect to the central galaxy.
|
Various events in the nature, economics and in other areas force us to
combine the study of extremes with regression and other methods. A useful tool
for reducing the role of nuisance regression, while we are interested in the
shape or tails of the basic distribution, is provided by the averaged
regression quantile and namely by the average extreme regression quantile. Both
are weighted means of regression quantile components, with weights depending on
the regressors. Our primary interest is the averaged extreme regression
quantile (AERQ), its structure, qualities and its applications, e.g. in
investigation of a conditional loss given a value exogenous economic and market
variables. AERQ has several interesting equivalent forms: While it is
originally defined as an optimal solution of a specific linear programming
problem, hence is a weighted mean of responses corresponding to the optimal
base of the pertaining linear program, we give another equivalent form as a
maximum residual of responses from a specific R-estimator of the slope
components of regression parameter. The latter form shows that while AERQ
equals to the maximum of some residuals of the responses, it has minimal
possible perturbation by the regressors. Notice that these finite-sample
results are true even for non-identically distributed model errors, e.g. under
heteroscedasticity. Moreover, the representations are formally true even when
the errors are dependent - this all provokes a question of the right
interpretation and of other possible applications.
|
The divergence-free time-independent velocity vector field has been
determined so as to maximise heat transfer between two parallel plates of a
constant temperature difference under the constraint of fixed total enstrophy.
The present variational problem is the same as that first formulated by
Hassanzadeh $\it et{\ }al$. (2014); however, a search range of optimal states
has been extended to a three-dimensional velocity field. The scaling of the
Nusselt number $Nu$ with the P\'eclet number $Pe$ (i.e., the square root of the
non-dimensionalised enstrophy with thermal diffusion timescale), $Nu\sim
Pe^{2/3}$, has been found in the three-dimensional optimal states,
corresponding to the asymptotic scaling with the Rayleigh number $Ra$, $Nu\sim
Ra^{1/2}$, in extremely-high-$Ra$ convective turbulence, and thus to the Taylor
energy dissipation law in high-Reynolds-number turbulence. At $Pe\sim10^{0}$, a
two-dimensional array of large-scale convection rolls provides maximal heat
transfer. A three-dimensional optimal solution emerges from bifurcation on the
two-dimensional solution branch at higher $Pe$. At $Pe\gtrsim10^{3}$, the
optimised velocity fields consist of convection cells with hierarchical
self-similar vortical structures, and the temperature fields exhibit a
logarithmic mean profile near the walls.
|
We present working notes on transfer learning with semi-supervised dataset
annotation for the BirdCLEF 2023 competition, focused on identifying African
bird species in recorded soundscapes. Our approach utilizes existing
off-the-shelf models, BirdNET and MixIT, to address representation and labeling
challenges in the competition. We explore the embedding space learned by
BirdNET and propose a process to derive an annotated dataset for supervised
learning. Our experiments involve various models and feature engineering
approaches to maximize performance on the competition leaderboard. The results
demonstrate the effectiveness of our approach in classifying bird species and
highlight the potential of transfer learning and semi-supervised dataset
annotation in similar tasks.
|
This paper develops a new global optimisation method that applies to a family
of criteria that are not entirely known. This family includes the criteria
obtained from the class of posteriors that have nor-malising constants that are
analytically not tractable. The procedure applies to posterior probability
densities that are continuously differen-tiable with respect to their
parameters. The proposed approach avoids the re-sampling needed for the
classical Monte Carlo maximum likelihood inference, while providing the missing
convergence properties of the ABC based methods. Results on simulated data and
real data are presented. The real data application fits an inhomogeneous area
interaction point process to cosmological data. The obtained results validate
two important aspects of the galaxies distribution in our Universe : proximity
of the galaxies from the cosmic filament network together with territorial
clustering at given range of interactions. Finally, conclusions and
perspectives are depicted.
|
We simulate the propagation of cosmic rays at ultra-high energies, $\gtrsim
10^{18}$ eV, in models of extragalactic magnetic fields in constrained
simulations of the local Universe. We use constrained initial conditions with
the cosmological magnetohydrodynamics code {\sc ENZO}. The resulting models of
the distribution of magnetic fields in the local Universe are used in the
\crpropa code to simulate the propagation of ultra-high energy cosmic rays. We
investigate the impact of six different magneto-genesis scenarios, both
primordial and astrophysical, on the propagation of cosmic rays over
cosmological distances. Moreover, we study the influence of different source
distributions around the Milky Way. Our study shows that different scenarios of
magneto-genesis do not have a large impact on the anisotropy measurements of
ultra-high energy cosmic rays. However, at high energies above the GZK-limit,
there is anisotropy caused by the distribution of nearby sources, independent
of the magnetic field model. This provides a chance to identify cosmic ray
sources with future full-sky measurements and high number statistics at the
highest energies. Finally, we compare our results to the dipole signal measured
by the Pierre Auger Observatory. All our source models and magnetic field
models could reproduce the observed dipole amplitude with a pure iron injection
composition. Our results indicate that the dipole is observed due to clustering
of secondary nuclei in direction of nearby sources of heavy nuclei. A light
injection composition is disfavoured by the non-observation of anisotropy at
energies of $4-8 \rm\ EeV$.
|
Hyperspectral Imaging (HSI) serves as an important technique in remote
sensing. However, high dimensionality and data volume typically pose
significant computational challenges. Band selection is essential for reducing
spectral redundancy in hyperspectral imagery while retaining intrinsic critical
information. In this work, we propose a novel hyperspectral band selection
model by decomposing the data into a low-rank and smooth component and a sparse
one. In particular, we develop a generalized 3D total variation (G3DTV) by
applying the $\ell_1^p$-norm to derivatives to preserve spatial-spectral
smoothness. By employing the alternating direction method of multipliers
(ADMM), we derive an efficient algorithm, where the tensor low-rankness is
implied by the tensor CUR decomposition. We demonstrate the effectiveness of
the proposed approach through comparisons with various other state-of-the-art
band selection techniques using two benchmark real-world datasets. In addition,
we provide practical guidelines for parameter selection in both noise-free and
noisy scenarios.
|
Nanolithography based on local oxidation with a scanning force microscope has
been performed on an undoped GaAs wafer and a Ga[Al]As heterostructure with an
undoped GaAs cap layer and a shallow two-dimensional electron gas. The oxide
growth and the resulting electronic properties of the patterned structures are
compared for constant and modulated voltage applied to the conductive tip of
the scanning force microscope. All the lithography has been performed in
non-contact mode. Modulating the applied voltage enhances the aspect ratio of
the oxide lines, which significantly strengthens the insulating properties of
the lines on GaAs. In addition, the oxidation process is found to be more
reliable and reproducible. Using this technique, a quantum point contact and a
quantum wire have been defined and the electronic stability, the confinement
potential and the electrical tunability are demonstrated to be similar to the
oxidation with constant voltage.
|
With the proliferating of wireless demands, wireless local area network
(WLAN) becomes one of the most important wireless networks. Network
intelligence is promising for the next generation wireless networks, captured
lots of attentions. Sensing is one efficient enabler to achieve network
intelligence since utilizing sensing can obtain diverse and valuable
non-communication information. Thus, integrating sensing and communications
(ISAC) is a promising technology for future wireless networks. Sensing assisted
communication (SAC) is an important branch of ISAC, but there are few related
works focusing on the systematical and comprehensive analysis on SAC in WLAN.
This article is the first work to systematically analyze SAC in the next
generation WLAN from the system simulation perspective. We analyze the
scenarios and advantages of SAC. Then, from system simulation perspective,
several sources of performance gain brought from SAC are proposed, i.e. beam
link failure, protocol overhead, and intra-physical layer protocol data unit
(intra-PPDU) performance decrease, while several important influencing factors
are described in detail. Performance evaluation is deeply analyzed and the
performance gain of the SAC in both living room and street canyon scenarios are
verified by system simulation. Finally, we provide our insights on the future
directions of SAC for the next generation WLAN.
|
We study a canonical C$^*$-algebra, $\mathcal{S}(\Gamma, \mu)$, that arises
from a weighted graph $(\Gamma, \mu)$, specific cases of which were previously
studied in the context of planar algebras. We discuss necessary and sufficient
conditions of the weighting which ensure simplicity and uniqueness of trace of
$\mathcal{S}(\Gamma, \mu)$, and study the structure of its positive cone. We
then study the $*$-algebra, $\mathcal{A}$, generated by the generators of
$\mathcal{S}(\Gamma, \mu)$, and use a free differential calculus and techniques
of Charlesworth and Shlyakhtenko, as well as Mai, Speicher, and Weber to show
that certain "loop" elements have no atoms in their spectral measure. After
modifying techniques of Shlyakhtenko and Skoufranis to show that self adjoint
elements $x \in M_{n}(\mathcal{A})$ have algebraic Cauchy transform, we explore
some applications to eigenvalues of polynomials in Wishart matrices and to
diagrammatic elements in von Neumann algebras initially considered by Guionnet,
Jones, and Shlyakhtenko.
|
We present a super-polynomial improvement in the precision scaling of quantum
simulations for coupled classical-quantum systems in this paper. Such systems
are found, for example, in molecular dynamics simulations within the
Born-Oppenheimer approximation. By employing a framework based on the
Koopman-von Neumann formalism, we express the Liouville equation of motion as
unitary dynamics and utilize phase kickback from a dynamical quantum simulation
to calculate the quantum forces acting on classical particles. This approach
allows us to simulate the dynamics of these particles without the overheads
associated with measuring gradients and solving the equations of motion on a
classical computer, resulting in a super-polynomial advantage at the price of
increased space complexity. We demonstrate that these simulations can be
performed in both microcanonical and canonical ensembles, enabling the
estimation of thermodynamic properties from the prepared probability density.
|
Variable OH/IR stars are Asymptotic Giant Branch (AGB) stars with an
optically thick circumstellar envelope that emit strong OH 1612 MHz emission.
They are commonly observed throughout the Galaxy but also in the LMC and SMC.
Hence, the precise inference of the distances of these stars will ultimately
result in better constraints on their mass range in different metallicity
environments. Through a multi-year long-term monitoring program at the Nancay
Radio telescope (NRT) and a complementary high-sensitivity mapping campaign at
the eMERLIN and JVLA to measure precisely the angular diameter of the
envelopes, we have been re-exploring distance determination through the
phase-lag method for a sample of stars, in order to refine the
poorly-constrained distances of some and infer the currently unknown distances
of others. We present here an update of this project.
|
In this work we report on the Landau gauge photon propagator computed for
pure gauge 4D compact QED in the confined and deconfined phases and for large
lattices volumes: $32^4$, $48^4$ and $96^4$. In the confined phase, compact QED
develops mass scales that render the propagator finite at all momentum scales
and no volume dependence is observed for the simulations performed.
Furthermore, for the confined phase the propagator is compatible with a Yukawa
massive type functional form. For the deconfined phase the photon propagator
seems to approach a free field propagator as the lattice volume is increased.
In both cases, we also investigate the static potential and the average value
of the number of Dirac strings in the gauge configurations $m$. In the confined
phase the mass gap translates into a linearly growing static potential, while
in the deconfined phase the static potential approaches a constant at large
separations. Results shows that $m$ is, at least, one order of magnitude larger
in the confined phase and confirm that the appearance of a confined phase is
connected with the topology of the gauge group.
|
It is argued that in the context of TeV gravity with large extra dimensions,
excited string states produced in colliders and in the interaction of cosmic
rays with the atmosphere may decay preferentially into invisible bulk modes,
rather than visible gauge fields on the brane. This contrast to the black hole
case comes about because of the absence of a relationship between physical size
and temperature for string ball states. We estimate the effect of this upon the
number of events predicted at cosmic ray observatories and colliders.
|
We describe a technique to emulate a two-level \PT-symmetric spin
Hamiltonian, replete with gain and loss, using only the unitary dynamics of a
larger quantum system. This we achieve by embedding the two-level system in
question in a subspace of a four-level Hamiltonian. Using an \textit{amplitude
recycling} scheme that couples the levels exterior to the \PT-symmetric
subspace, we show that it is possible to emulate the desired behaviour of the
\PT-symmetric Hamiltonian without depleting the exterior, reservoir levels. We
are thus able to extend the emulation time indefinitely, despite the
non-unitary \PT dynamics. We propose a realistic experimental implementation
using dynamically decoupled magnetic sublevels of ultracold atoms.
|
A theory of topological gravity is a homotopy-theoretic representation of the
Segal-Tillmann topologification of a two-category with cobordisms as morphisms.
This note describes a relatively accessible example of such a thing, suggested
by the wall-crossing formulas of Donaldson theory.
|
The observation of strongly-correlated states in moir\'e systems has renewed
the conceptual interest in magnetic systems with higher SU(4) spin symmetry,
e.g. to describe Mott insulators where the local moments are coupled
spin-valley degrees of freedom. Here, we discuss a numerical renormalization
group scheme to explore the formation of spin-valley ordered and unconventional
spin-valley liquid states at zero temperature based on a pseudo-fermion
representation. Our generalization of the conventional pseudo-fermion
functional renormalization group approach for $\mathfrak{su}$(2) spins is
capable of treating diagonal and off-diagonal couplings of generic spin-valley
exchange Hamiltonians in the self-conjugate representation of the
$\mathfrak{su}$(4) algebra. To achieve proper numerical efficiency, we derive a
number of symmetry constraints on the flow equations that significantly limit
the number of ordinary differential equations to be solved. As an example
system, we investigate a diagonal SU(2)$_{\textrm{spin}}$ $\otimes$
U(1)$_{\textrm{valley}}$ model on the triangular lattice which exhibits a rich
phase diagram of spin and valley ordered phases.
|
A significant challenge in the field of object detection lies in the system's
performance under non-ideal imaging conditions, such as rain, fog, low
illumination, or raw Bayer images that lack ISP processing. Our study
introduces "Feature Corrective Transfer Learning", a novel approach that
leverages transfer learning and a bespoke loss function to facilitate the
end-to-end detection of objects in these challenging scenarios without the need
to convert non-ideal images into their RGB counterparts. In our methodology, we
initially train a comprehensive model on a pristine RGB image dataset.
Subsequently, non-ideal images are processed by comparing their feature maps
against those from the initial ideal RGB model. This comparison employs the
Extended Area Novel Structural Discrepancy Loss (EANSDL), a novel loss function
designed to quantify similarities and integrate them into the detection loss.
This approach refines the model's ability to perform object detection across
varying conditions through direct feature map correction, encapsulating the
essence of Feature Corrective Transfer Learning. Experimental validation on
variants of the KITTI dataset demonstrates a significant improvement in mean
Average Precision (mAP), resulting in a 3.8-8.1% relative enhancement in
detection under non-ideal conditions compared to the baseline model, and a less
marginal performance difference within 1.3% of the mAP@[0.5:0.95] achieved
under ideal conditions by the standard Faster RCNN algorithm.
|
Comment on "Critical states and fractal attractors in fractal tongues:
Localization in the Harper map" [Phys. Rev. E64 (2001) 045204]
|
It is pointed out that simulation computation of energy performed so far
cannot be used to decide if the ground state of solid 4He has the number of
lattice sites equal to the number of atoms (commensurate state) or if it is
different (incommensurate state). The best variational wave function, a shadow
wave function, gives an incommensurate state but the equilibrium concentration
of vacancies remains to be determined. In order to investigate the presence of
a supersolid phase we have computed the one--body density matrix in solid 4He
for the incommensurate state by means of the exact Shadow Path Integral Ground
State projector method. We find a vacancy induced Bose Einstein condensation of
about 0.23 atoms per vacancy at a pressure of 54 bar. This means that bulk
solid 4He is supersolid at low enough temperature if the exact ground state is
incommensurate.
|
Small operators who take part in secondary wireless spectrum markets
typically have strict budget limits. In this paper, we study the bidding
problem of a budget constrained operator in repeated secondary spectrum
auctions. In existing truthful auctions, truthful bidding is the optimal
strategy of a bidder. However, budget limits impact bidding behaviors and make
bidding decisions complicated, since bidders may behave differently to avoid
running out of money. We formulate the problem as a dynamic auction game
between operators, where knowledge of other operators is limited due to the
distributed nature of wireless networks/markets. We first present a Markov
Decision Process (MDP) formulation of the problem and characterize the optimal
bidding strategy of an operator, provided that opponents' bids are i.i.d. Next,
we generalize the formulation to a Markov game that, in conjunction with
model-free reinforcement learning approaches, enables an operator to make
inferences about its opponents based on local observations. Finally, we present
a fully distributed learning-based bidding algorithm which relies only on local
information. Our numerical results show that our proposed learning-based
bidding results in a better utility than truthful bidding.
|
Hot Jupiters are rarely accompanied by other planets within a factor of a few
in orbital distance. Previously, only two such systems have been found. Here,
we report the discovery of a third system using data from the Transiting
Exoplanet Survey Satellite (TESS). The host star, TOI-1130, is an 11th
magnitude K-dwarf in the Gaia G band. It has two transiting planets: a
Neptune-sized planet ($3.65\pm 0.10$ $R_E$) with a 4.1-day period, and a hot
Jupiter ($1.50^{+0.27}_{-0.22}$ $R_J$) with an 8.4-day period. Precise
radial-velocity observations show that the mass of the hot Jupiter is
$0.974^{+0.043}_{-0.044}$ $M_J$. For the inner Neptune, the data provide only
an upper limit on the mass of 0.17 $M_J$ (3$\sigma$). Nevertheless, we are
confident the inner planet is real, based on follow-up ground-based photometry
and adaptive optics imaging that rule out other plausible sources of the TESS
transit signal. The unusual planetary architecture of and the brightness of the
host star make TOI-1130 a good test case for planet formation theories, and an
attractive target for future spectroscopic observations.
|
Intra-day economic dispatch of an integrated microgrid is a fundamental
requirement to integrate distributed generators. The dynamic energy flows in
cogeneration units present challenges to the energy management of the
microgrid. In this paper, a novel approximate dynamic programming (ADP)
approach is proposed to solve this problem based on value function
approximation, which is distinct with the consideration of the dynamic process
constraints of the combined-cycle gas turbine (CCGT) plant. First, we
mathematically formulate the multi-time periods decision problem as a
finite-horizon Markov decision process. To deal with the thermodynamic process,
an augmented state vector of CCGT is introduced. Second, the proposed VFA-ADP
algorithm is employed to derive the near-optimal real-time operation
strategies. In addition, to guarantee the monotonicity of piecewise linear
function, we apply the SPAR algorithm in the update process. To validate the
effectiveness of the proposed method, we conduct experiments with comparisons
to some traditional optimization methods. The results indicate that our
proposed ADP method achieves better performance on the economic dispatch of the
microgrid.
|
Ensemble methods are a cornerstone of modern machine learning. The
performance of an ensemble depends crucially upon the level of diversity
between its constituent learners. This paper establishes a connection between
diversity and degrees of freedom (i.e. the capacity of the model), showing that
diversity may be viewed as a form of inverse regularisation. This is achieved
by focusing on a previously published algorithm Negative Correlation Learning
(NCL), in which model diversity is explicitly encouraged through a diversity
penalty term in the loss function. We provide an exact formula for the
effective degrees of freedom in an NCL ensemble with fixed basis functions,
showing that it is a continuous, convex and monotonically increasing function
of the diversity parameter. We demonstrate a connection to Tikhonov
regularisation and show that, with an appropriately chosen diversity parameter,
an NCL ensemble can always outperform the unregularised ensemble in the
presence of noise. We demonstrate the practical utility of our approach by
deriving a method to efficiently tune the diversity parameter. Finally, we use
a Monte-Carlo estimator to extend the connection between diversity and degrees
of freedom to ensembles of deep neural networks.
|
We explore general scalar-tensor models in the presence of a kinetic mixing
between matter and the scalar field, which we call Kinetic Matter Mixing. In
the frame where gravity is de-mixed from the scalar this is due to disformal
couplings of matter species to the gravitational sector, with disformal
coefficients that depend on the gradient of the scalar field. In the frame
where matter is minimally coupled, it originates from the so-called beyond
Horndeski quadratic Lagrangian. We extend the Effective Theory of Interacting
Dark Energy by allowing disformal coupling coefficients to depend on the
gradient of the scalar field as well. In this very general approach, we derive
the conditions to avoid ghost and gradient instabilities and we define Kinetic
Matter Mixing independently of the frame metric used to described the action.
We study its phenomenological consequences for a $\Lambda$CDM background
evolution, first analytically on small scales. Then, we compute the matter
power spectrum and the angular spectra of the CMB anisotropies and the CMB
lensing potential, on all scales. We employ the public version of COOP, a
numerical Einstein-Boltzmann solver that implements very general scalar-tensor
modifications of gravity. Rather uniquely, Kinetic Matter Mixing weakens
gravity on short scales, predicting a lower $\sigma_8$ with respect to the
$\Lambda$CDM case. We propose this as a possible solution to the tension
between the CMB best-fit model and low-redshift observables.
|
Spectral properties of Schr\"odinger operators on compact metric graphs are
studied and special emphasis is put on differences in the spectral behavior
between different classes of vertex conditions. We survey recent results
especially for $\delta$ and $\delta'$ couplings and demonstrate the spectral
properties on many examples. Amongst other things, properties of the ground
state eigenvalue and eigenfunction and the spectral behavior under various
perturbations of the metric graph or the vertex conditions are considered.
|
The successive ionization potentials (IPs) and electron affinities (EAs) for
superheavy elements with $111 \leq Z \leq 114$, namely, Rg, Cn, Nh, and Fl are
reexamined using the relativistic Fock-space coupled-cluster method with
nonperturbative single (S), double (D), and triple (T) cluster amplitudes
(FS-CCSDT). For the most of considered quantities, the triple-amplitude
contributions turn out to be important. The Breit and frequency-dependent Breit
corrections are evaluated by means of the configuration-interaction method. The
quantum-electrodynamics corrections to the IPs and EAs are taken into account
within the model-QED-operator approach. The obtained results are within 0.10 eV
uncertainty.
|
The evolutionary persistence of symbiotic associations is a puzzle.
Adaptation should eliminate cooperative traits if it is possible to enjoy the
advantages of cooperation without reciprocating - a facet of cooperation known
in game theory as the Prisoner's Dilemma. Despite this barrier, symbioses are
widespread, and may have been necessary for the evolution of complex life. The
discovery of strategies such as tit-for-tat has been presented as a general
solution to the problem of cooperation. However, this only holds for
within-species cooperation, where a single strategy will come to dominate the
population. In a symbiotic association each species may have a different
strategy, and the theoretical analysis of the single species problem is no
guide to the outcome. We present basic analysis of two-species cooperation and
show that a species with a fast adaptation rate is enslaved by a slowly
evolving one. Paradoxically, the rapidly evolving species becomes highly
cooperative, whereas the slowly evolving one gives little in return. This helps
understand the occurrence of endosymbioses where the host benefits, but the
symbionts appear to gain little from the association.
|
The decision whether a measured distribution complies with an
equidistribution is a central element of many biostatistical methods. High
throughput differential expression measurements, for instance, necessitate to
judge possible over-representation of genes. The reliability of this judgement,
however, is strongly affected when rarely expressed genes are pooled. We
propose a method that can be applied to frequency ranked distributions and that
yields a simple but efficient criterion to assess the hypothesis of
equiprobable expression levels. By applying our technique to surrogate data we
exemplify how the decision criterion can differentiate between a true
equidistribution and a triangular distribution. The distinction succeeds even
for small sample sizes where standard tests of significance (e.g. chi^2) fail.
Our method will have a major impact on several problems of computational
biology where rare events baffle a reliable assessment of frequency
distributions.
|
We investigate double finger gate (DFG) controlled spin-resolved resonant
transport properties in an n-type quantum channel with a Rashba-Zeeman (RZ)
subband energy gap. By appropriately tuning the DFG in the strong Rashba
coupling regime, resonant state structures in conductance can be found that is
sensitive to the length of the DFG system. Furthermore, a hole-like bound state
feature below the RZ gap and an electron-like quasi-bound state feature at the
threshold of the upper spin branch can be found that is insensitive to the
length of the DFG system.
|
Gravitational waves from coalescences of neutron stars or stellar-mass black
holes into intermediate-mass black holes (IMBHs) of $\gtrsim 100$ solar masses
represent one of the exciting possible sources for advanced gravitational-wave
detectors. These sources can provide definitive evidence for the existence of
IMBHs, probe globular-cluster dynamics, and potentially serve as tests of
general relativity. We analyse the accuracy with which we can measure the
masses and spins of the IMBH and its companion in intermediate-mass ratio
coalescences. We find that we can identify an IMBH with a mass above $100 ~
M_\odot$ with $95\%$ confidence provided the massive body exceeds $130 ~
M_\odot$. For source masses above $\sim200 ~ M_\odot$, the best measured
parameter is the frequency of the quasi-normal ringdown. Consequently, the
total mass is measured better than the chirp mass for massive binaries, but the
total mass is still partly degenerate with spin, which cannot be accurately
measured. Low-frequency detector sensitivity is particularly important for
massive sources, since sensitivity to the inspiral phase is critical for
measuring the mass of the stellar-mass companion. We show that we can
accurately infer source parameters for cosmologically redshifted signals by
applying appropriate corrections. We investigate the impact of uncertainty in
the model gravitational waveforms and conclude that our main results are likely
robust to systematics.
|
We determine the contribution of nontrivial vacuum (topological) excitations,
more specifically vortex--strings of the Abelian Higgs model in 3+1 dimensions,
to the functional partition function. By expressing the original action in
terms of dual transformed fields we make explicit in the equivalent action the
contribution of the vortex--strings excitations of the model. The effective
potential of an appropriately defined local vacuum expectation value of the
vortex--string field in the dual transformed action is then evaluated both at
zero and finite temperatures and its properties discussed in the context of the
finite temperature phase transition.
|
The end-to-end learning pipeline is gradually creating a paradigm shift in
the ongoing development of highly autonomous vehicles, largely due to advances
in deep learning, the availability of large-scale training datasets, and
improvements in integrated sensor devices. However, a lack of interpretability
in real-time decisions with contemporary learning methods impedes user trust
and attenuates the widespread deployment and commercialization of such
vehicles. Moreover, the issue is exacerbated when these cars are involved in or
cause traffic accidents. Such drawback raises serious safety concerns from
societal and legal perspectives. Consequently, explainability in end-to-end
autonomous driving is essential to build trust in vehicular automation.
However, the safety and explainability aspects of end-to-end driving have
generally been investigated disjointly by researchers in today's state of the
art. This survey aims to bridge the gaps between these topics and seeks to
answer the following research question: When and how can explanations improve
safety of end-to-end autonomous driving? In this regard, we first revisit
established safety and state-of-the-art explainability techniques in end-to-end
driving. Furthermore, we present three critical case studies and show the
pivotal role of explanations in enhancing self-driving safety. Finally, we
describe insights from empirical studies and reveal potential value,
limitations, and caveats of practical explainable AI methods with respect to
their safety assurance in end-to-end autonomous driving.
|
The scale factor $\sigma_{eff}$, which characterizes double parton collisions
in high energy hadron interactions, is a direct manifestation of the
distribution of the interacting partons in transverse space, in such a way that
different distributions give rise to different values of $\sigma_{eff}$ in
different double parton collision processes. We work out the value of the scale
factor in a few reactions of interest, in a correlated model of the
multi-parton density of the proton recently proposed.
|
Two-stream architecture have shown strong performance in video classification
task. The key idea is to learn spatio-temporal features by fusing convolutional
networks spatially and temporally. However, there are some problems within such
architecture. First, it relies on optical flow to model temporal information,
which are often expensive to compute and store. Second, it has limited ability
to capture details and local context information for video data. Third, it
lacks explicit semantic guidance that greatly decrease the classification
performance. In this paper, we proposed a new two-stream based deep framework
for video classification to discover spatial and temporal information only from
RGB frames, moreover, the multi-scale pyramid attention (MPA) layer and the
semantic adversarial learning (SAL) module is introduced and integrated in our
framework. The MPA enables the network capturing global and local feature to
generate a comprehensive representation for video, and the SAL can make this
representation gradually approximate to the real video semantics in an
adversarial manner. Experimental results on two public benchmarks demonstrate
our proposed methods achieves state-of-the-art results on standard video
datasets.
|
Memory-assisted measurement-device-independent quantum key distribution
(MA-MDI-QKD) is a promising scheme that aims to improve the
rate-versus-distance behavior of a QKD system by using the state-of-the-art
devices. It can be seen as a bridge between current QKD links to quantum
repeater based networks. While, similar to quantum repeaters, MA-MDI-QKD relies
on quantum memory (QM) units, the requirements for such QMs are less demanding
than that of probabilistic quantum repeaters. Here, we present a variant of
MA-MDI-QKD structure that relies on only a single physical QM: a
nitrogen-vacancy center embedded into a cavity where its electronic spin
interacts with photons and its nuclear spin is used for storage. This enables
us to propose a simple but efficient MA-MDI-QKD scheme resilient to memory
errors and capable of beating, in terms of rate and reach, existing QKD
demonstrations. We also show how we can extend this setup to a quantum repeater
system, reaching, thus, larger distances.
|
Entanglement generation in microcavity exciton-polaritons is an interesting
application of the peculiar properties of these half-light/half-matter
quasiparticles. In this paper we theoretically investigate their luminescence
dynamics and entanglement formation in single, double, and triple cavities. We
derive general expressions and selection rules for polariton-polariton
scattering. We evaluate a number of possible parametric scattering schemes in
terms of entanglement, and identify the ones that are experimentally most
promising.
|
The turbulence induced decay of orbital angular momentum (OAM) entanglement
between two photons is investigated numerically and experimentally. To compare
our results with previous work, we simulate the turbulent atmosphere with a
single phase screen based on the Kolmogorov theory of turbulence. We consider
two different scenarios: in the first only one of the two photons propagates
through turbulence, and in the second both photons propagate through
uncorrelated turbulence. Comparing the entanglement evolution for different OAM
values, we found the entanglement to be more robust in turbulence for higher
OAM values. We derive an empirical formula for the distance scale at which
entanglement decays in term of the scale parameters and the OAM value.
|
This paper investigates battery preheating before fast charging, for a
battery electric vehicle (BEV) driving in a cold climate. To prevent the
battery from performance degradation at low temperatures, a thermal management
(TM) system has been considered, including a high-voltage coolant heater (HVCH)
for the battery and cabin compartment heating. Accordingly, an optimal control
problem (OCP) has been formulated in the form of a nonlinear program (NLP),
aiming at minimising the total energy consumption of the battery. The main
focus here is to develop a computationally efficient approach, mimicking the
optimal preheating behavior without a noticeable increase in the total energy
consumption. The proposed algorithm is simple enough to be implemented in a
low-level electronic control unit of the vehicle, by eliminating the need for
solving the full NLP in the cost of only 1Wh increase in the total energy
consumption.
|
We provide experimental evidence that the upper limit of ~110 K commonly
observed for the Curie temperature T_C of Ga(1-x)Mn(x)As is caused by the
Fermi-level-induced hole saturation. Ion channeling, electrical and
magnetization measurements on a series of Ga(1-x-y)Mn(x)Be(y)As layers show a
dramatic increase of the concentration of Mn interstitials accompanied by a
reduction of T_C with increasing Be concentration, while the free hole
concentration remains relatively constant at ~5x10^20 cm^-3. These results
indicate that the concentrations of free holes and ferromagnetically active Mn
spins are governed by the position of the Fermi level, which controls the
formation energy of compensating interstitial Mn donors.
|
We present a fourth catalog of HI sources from the Arecibo Legacy Fast ALFA
(ALFALFA) Survey. We report 541 detections over 136 deg2, within the region of
the sky having 22h < R.A. < 03h and 24 deg < Dec. < 26 deg . This complements a
previous catalog in the region 26 deg < Dec. < 28 deg (Saintonge et al. 2008).
We present here the detections falling into three classes: (a) extragalactic
sources with S/N > 6.5, where the reliability of the catalog is better than
95%; (b) extragalactic sources 5.0 < S/N < 6.5 and a previously measured
optical redshift that corroborates our detection; or (c) High Velocity Clouds
(HVCs), or subcomponents of such clouds, in the periphery of the Milky Way. Of
the 541 objects presented here, 90 are associated with High Velocity Clouds,
while the remaining 451 are identified as extragalactic objects. Optical
counterparts have been matched with all but one of the extragalactic objects.
|
We introduce a novel rule-based approach for handling regression problems.
The new methodology carries elements from two frameworks: (i) it provides
information about the uncertainty of the parameters of interest using Bayesian
inference, and (ii) it allows the incorporation of expert knowledge through
rule-based systems. The blending of those two different frameworks can be
particularly beneficial for various domains (e.g. engineering), where, even
though the significance of uncertainty quantification motivates a Bayesian
approach, there is no simple way to incorporate researcher intuition into the
model. We validate our models by applying them to synthetic applications: a
simple linear regression problem and two more complex structures based on
partial differential equations. Finally, we review the advantages of our
methodology, which include the simplicity of the implementation, the
uncertainty reduction due to the added information and, in some occasions, the
derivation of better point predictions, and we address limitations, mainly from
the computational complexity perspective, such as the difficulty in choosing an
appropriate algorithm and the added computational burden.
|
The thermal conductivity of the layered s-wave superconductor NbSe_2 was
measured down to T_c/100 throughout the vortex state. With increasing field, we
identify two regimes: one with localized states at fields very near H_c1 and
one with highly delocalized quasiparticle excitations at higher fields. The two
associated length scales are most naturally explained as multi-band
superconductivity, with distinct small and large superconducting gaps on
different sheets of the Fermi surface.
|
The choice of making an intervention depends on its potential benefit or harm
in comparison to alternatives. Estimating the likely outcome of alternatives
from observational data is a challenging problem as all outcomes are never
observed, and selection bias precludes the direct comparison of differently
intervened groups. Despite their empirical success, we show that algorithms
that learn domain-invariant representations of inputs (on which to make
predictions) are often inappropriate, and develop generalization bounds that
demonstrate the dependence on domain overlap and highlight the need for
invertible latent maps. Based on these results, we develop a deep kernel
regression algorithm and posterior regularization framework that substantially
outperforms the state-of-the-art on a variety of benchmarks data sets.
|
Within multi-Higgs-doublet models, one can impose symmetries on the Higgs
potential, either discrete or continuous, that mix several doublets. In
two-Higgs-doublet model any such symmetry can be conserved or spontaneously
violated after the electroweak symmetry breaking (EWSB), depending on the
coefficients of the potential. With more than two doublets, there exist
symmetries which are always spontaneously violated after EWSB. We discuss the
origin of this phenomenon and show its similarity to geometric frustration in
condensed-matter physics.
|
We study the challenging problem of recovering detailed motion from a single
motion-blurred image. Existing solutions to this problem estimate a single
image sequence without considering the motion ambiguity for each region.
Therefore, the results tend to converge to the mean of the multi-modal
possibilities. In this paper, we explicitly account for such motion ambiguity,
allowing us to generate multiple plausible solutions all in sharp detail. The
key idea is to introduce a motion guidance representation, which is a compact
quantization of 2D optical flow with only four discrete motion directions.
Conditioned on the motion guidance, the blur decomposition is led to a
specific, unambiguous solution by using a novel two-stage decomposition
network. We propose a unified framework for blur decomposition, which supports
various interfaces for generating our motion guidance, including human input,
motion information from adjacent video frames, and learning from a video
dataset. Extensive experiments on synthesized datasets and real-world data show
that the proposed framework is qualitatively and quantitatively superior to
previous methods, and also offers the merit of producing physically plausible
and diverse solutions. Code is available at
https://github.com/zzh-tech/Animation-from-Blur.
|
In this paper we describe all possible reduced complete intersection sets of
points on Veronese surfaces. We formulate a conjecture for the general case of
complete intersection subvarieties of any dimension and we prove it in the case
of the quadratic Veronese threefold. Our main tool is an effective
characterization of all possible Hilbert functions of reduced subvarieties of
Veronese surfaces.
|
Linear thresholding systems have been used as a model of neural activation
and more recently proposed as a model of gene regulation. Here we exhibit
linear thresholding systems whose dynamics produce surprisingly long cycles.
|
Hereditary hemolytic anemias are genetic disorders that affect the shape and
density of red blood cells. Genetic tests currently used to diagnose such
anemias are expensive and unavailable in the majority of clinical labs. Here,
we propose a method for identifying hereditary hemolytic anemias based on a
standard biochemistry method, called Percoll gradient, obtained by centrifuging
a patient's blood. Our hybrid approach consists on using spatial data-driven
features, extracted with a convolutional neural network and spectral
handcrafted features obtained from fast Fourier transform. We compare late and
early feature fusion with AlexNet and VGG16 architectures. AlexNet with late
fusion of spectral features performs better compared to other approaches. We
achieved an average F1-score of 88% on different classes suggesting the
possibility of diagnosing of hereditary hemolytic anemias from Percoll
gradients. Finally, we utilize Grad-CAM to explore the spatial features used
for classification.
|
This article presents an approach to encode Linear Temporal Logic (LTL)
Specifications into a Mixed Integer Quadratically Constrained Quadratic Program
(MIQCQP) footstep planner. We propose that the integration of LTL
specifications into the planner not only facilitates safe and desirable
locomotion between obstacle-free regions, but also provides a rich language for
high-level reasoning in contact planning. Simulations of the footstep planner
in a 2D environment satisfying encoded LTL specifications demonstrate the
results of this research.
|
In this paper, we provide novel definitions of clustering coefficient for
weighted and directed multilayer networks. We extend in the multilayer
theoretical context the clustering coefficients proposed in the literature for
weighted directed monoplex networks. We quantify how deeply a node is involved
in a choesive structure focusing on a single node, on a single layer or on the
entire system. The coefficients convey several characteristics inherent to the
complex topology of the multilayer network. We test their effectiveness
applying them to a particularly complex structure such as the international
trade network. The trade data integrate different aspects and they can be
described by a directed and weighted multilayer network, where each layer
represents import and export relationships between countries for a given
sector. The proposed coefficients find successful application in describing the
interrelations of the trade network, allowing to disentangle the effects of
countries and sectors and jointly consider the interactions between them.
|
Distributions of Monge type are a class of strongly regular
bracket-generating distributions introduced by I. Anderson, Zh. Nie and P.
Nurowski. Their symbol algebras prolong to simple graded Lie algebras, thus
allowing one to associate a parabolic geometry to any given Monge distribution.
This article is devoted to the classification problem for homogeneous models of
Monge distributions of type C3 in dimension eight, and is complementary to a
paper by I. Anderson and P. Nurowski. The general classification algorithm, as
well as most of its application to the particular problem, are joint work with
Ian Anderson.
|
In artificial spin ice systems, an interplay of defects and dipolar
interactions is expected to play important roles in stabilizing different
collective magnetic states. In this work, we investigated the magnetization
reversal of individual defective square artificial spin ice vertices where
defects break four-fold rotational symmetry of the system. By varying the angle
between the applied field and the geometrical axis of the vertices, we observe
a change in energy landscape of the system resulting into the stabilization of
collective low-energy magnetic states. We also observe that by changing the
angle, it is possible to access different vertex configurations. Micromagnetic
simulations are performed for varying angle as well as external field, the
results of which are consistent with the experimental data.
|
In this paper, we consider the fractional elliptic equation \begin{align*}
\left\{\begin{aligned} &(-\Delta)^s u-\mu\frac{u}{|x|^{2s}} =
\frac{|u|^{2_s^\ast (\alpha)-2}u}{|x|^{\alpha}} + f(x,u), && \mbox{in} \
\Omega,\\ &u=0, && \mbox{in} \ \mathbb{R}^{n}\backslash \ \Omega,
\end{aligned}\right. \end{align*} where $\Omega\subset R^n$ is a smooth bounded
domain, $0\in\Omega$, $0<s<1$, $0<\alpha<2s<n$,
$2_{s}^{\ast}(\alpha)=\frac{2(n-\alpha)}{n-2s}$. Under some assumptions on
$\mu$ and $f$, we obtain the existence of nonnegative solutions.
|
It is shown that instantons provide a natural mechanism to explain an unusual
azimuthal dependence of the production of the even-parity glueball candidates
in central pp collision. A different azimuthal dependence for instanton-induced
production of the odd-parity glueballs is predicted.
|
Let $\sigma_1$ and $\sigma_2$ be commuting involutions of a semisimple
algebraic group $G$. This yields a $Z_2\times Z_2$-grading of $\g=\Lie(G)$,
$\g=\bigoplus_{i,j=0,1}\g_{ij}$, and we study invariant-theoretic aspects of
this decomposition. Let $\g<\sigma_1>$ be the $Z_2$-contraction of $\g$
determined by $\sigma_1$. Then both $\sigma_2$ and $\sigma_3:=\sigma_1\sigma_2$
remain involutions of the non-reductive Lie algebra $\g<\sigma_1>$. The
isotropy representations related to $(\g<\sigma_1>, \sigma_2)$ and
$(\g<\sigma_1>, \sigma_3)$ are degenerations of the isotropy representations
related to $(\g, {\sigma_2})$ and $(\g, {\sigma_3})$, respectively. We show
that these degenerated isotropy representations retain many good properties.
For instance, they always have a generic stabiliser and their algebras of
invariants are often polynomial. We also develop some theory on Cartan
subspaces for various $Z_2$-gradings associated with the $Z_2\times
Z_2$-grading of $\g$.
|
In this article, we suggest a categorification procedure in order to capture
an analogy between Crystalline Grothendieck-Lefschetz trace formula and the
cyclotomic trace map $K\rightarrow TC$ from the algebraic $K$-theory to the
topological cyclic homology $TC$. First, we categorify the category of schemes
to the $(2, \infty)$-category of noncommuatative schemes a la Kontsevich. This
gives a categorification of the set of rational points of a scheme. Then, we
categorify the Crystalline Grothendieck-Lefschetz trace formula and find an
analogue to the Crystalline cohomology in the setting of noncommuative schemes
over $\mathbf{F}_{p}$. Our analogy suggests the existence of a categorification
of the $l$-adic cohomology trace formula in the noncommutative setting for
$l\neq p$. Finally, we write down the corresponding dictionary.
|
In manufacturing, the increasing involvement of autonomous robots in
production processes poses new challenges on the production management. In this
paper we report on the usage of Optimization Modulo Theories (OMT) to solve
certain multi-robot scheduling problems in this area. Whereas currently
existing methods are heuristic, our approach guarantees optimality for the
computed solution. We do not only present our final method but also its
chronological development, and draw some general observations for the
development of OMT-based approaches.
|
We propose a method for performing software pipelining on quantum for-loop
programs, exploiting parallelism in and across iterations. We redefine concepts
that are useful in program optimization, including array aliasing, instruction
dependency and resource conflict, this time in optimization of quantum
programs. Using the redefined concepts, we present a software pipelining
algorithm exploiting instruction-level parallelism in quantum loop programs.
The optimization method is then evaluated on some test cases, including popular
applications like QAOA, and compared with several baseline results. The
evaluation results show that our approach outperforms loop optimizers
exploiting only in-loop optimization chances by reducing total depth of the
loop program to close to the optimal program depth obtained by full loop
unrolling, while generating much smaller code in size. This is the first step
towards optimization of a quantum program with such loop control flow as far as
we know.
|
The topological phases of matter provide the opportunity to observe many
exotic properties, like the existence of two dimensional topological surface
states in the form of Dirac cone in topological insulators, chiral transport
through open Fermi arc in Weyl semimetals etc. However, these properties can
only affect the transport characteristics and therefore can be useful for
applications only if the topological phenomena occur near the Fermi level.
CaAgAs is a promising candidate, wherein the ab-initio calculations predict
line-node at the Fermi level which on including spin-orbit coupling transforms
into a topological insulator. In this report, we study the electronic structure
of CaAgAs with angle resolved photoemission spectroscopy (ARPES), ab-initio
calculations and transport measurements. The ARPES results show that the bulk
valence band crosses the Fermi energy at gamma-point and the band dispersion
matches the ab-initio calculations closely on shifting the Fermi energy by -0.5
eV. ARPES results are in good agreement with our transport measurements which
show abundant p-type carriers.
|
Motivated by the successful idea of using weakly-coupled quantum electronic
wires to realize the quantum Hall effects and the quantum spin Hall effects, we
theoretically construct two systems composed of weakly-coupled quantum spin
chains, which can exhibit spin analogues of superconductivity and the integer
quantum Hall effect. Specifically, a certain bilayer of two arrays of
interacting spin chains is mapped, via the Jordan-Wigner transformation, to a
negative-$U$ Hubbard model that exhibits superconductivity. In addition, an
array of spin-orbit-coupled spin chains in the presence of an suitable external
magnetic field is transformed to an array of quantum wires that exhibits the
integer quantum Hall effect. The resultant spin superconductivity and spin
integer quantum Hall effect can be characterized by their ability to transport
spin without any resistance.
|
We study the evaporation process of a 2D black hole in thermal equilibrium
when the ingoing radiation is switched off suddenly. We also introduce global
symmetries of generic 2D dilaton gravity models which generalize the extra
symmetry of the CGHS model.
|
The mechanical behaviour of two types of pasta (noodles and bucatini) was
studied in a cantilever-loaded-at-the-end experimental setup. One end of each
pasta was fixed while the other end was submitted to forces perpendicular to
the line determined by the pasta when undeflected. Elastic curves were studied,
resulting in values of $E=2.60$ GPa and $E=2.26$ GPa for the Young's modulus of
bucatini and noodles respectively. The relation coming from small slopes
approximation between the free end's displacement and the load was analyzed,
resulting in values of $E=2.33$ GPa and $E=2.44$ GPa for the Young's modulus of
bucatini and noodles respectively. Mechanical hysteresis was found in the
pasta, resulting in a small deformation. This experiment can be done with low
cost materials and it is a good first introduction to some basic concepts of
elasticity for mechanics courses.
|
We consider a discrete, non-Hermitian random matrix model, which can be
expressed as a shift of a rank-one perturbation of an anti-symmetric matrix. We
show that, asymptotically almost surely, the real parts of the eigenvalues of
the non-Hermitian matrix around any fixed index remain interlaced with those of
the anti-symmetric matrix. Along the way, we show that some tools recently
developed to study the eigenvalue distributions of Hermitian matrices extend to
the anti-symmetric setting.
|
Short-term demand forecasting models commonly combine convolutional and
recurrent layers to extract complex spatiotemporal patterns in data. Long-term
histories are also used to consider periodicity and seasonality patterns as
time series data. In this study, we propose an efficient architecture,
Temporal-Guided Network (TGNet), which utilizes graph networks and
temporal-guided embedding. Graph networks extract invariant features to
permutations of adjacent regions instead of convolutional layers.
Temporal-guided embedding explicitly learns temporal contexts from training
data and is substituted for the input of long-term histories from days/weeks
ago. TGNet learns an autoregressive model, conditioned on temporal contexts of
forecasting targets from temporal-guided embedding. Finally, our model achieves
competitive performances with other baselines on three spatiotemporal demand
dataset from real-world, but the number of trainable parameters is about 20
times smaller than a state-of-the-art baseline. We also show that
temporal-guided embedding learns temporal contexts as intended and TGNet has
robust forecasting performances even to atypical event situations.
|
Consider an inextensible closed filament immersed in a 2D Stokes fluid. Given
a force density $\mathbf{F}$ defined on this filament, we consider the problem
of determining the tension $\sigma$ on this filament that ensures the filament
is inextensible. This is a subproblem of dynamic inextensible vesicle and
membrane problems, which appear in engineering and biological applications. We
study the well-posedness and regularity properties of this problem in H\"older
spaces. We find that the tension determination problem admits a unique solution
if and only if the closed filament is {\em not} a circle. Furthermore, we show
that the tension $\sigma$ gains one derivative with respect to the imposed line
force density $\mathbf{F}$, and show that the tangential and normal components
of $\mathbf{F}$ affect the regularity of $\sigma$ in different ways. We also
study the near singularity of the tension determination problem as the
interface approaches a circle, and verify our analytical results against
numerical experiment.
|
In serverless computing, applications are executed under lightweight
virtualization and isolation environments, such as containers or micro virtual
machines. Typically, their memory allocation is set by the user before
deployment. All other resources, such as CPU, are allocated by the provider
statically and proportionally to memory allocations. This contributes to either
under-utilization or throttling. The former significantly impacts the provider,
while the latter impacts the client. To solve this problem and accommodate both
clients and providers, a solution is dynamic CPU allocation achieved through
autoscaling. Autoscaling has been investigated for long-running applications
using history-based techniques and prediction. However, serverless applications
are short-running workloads, where such techniques are not well suited. In this
paper, we investigate tiny autoscalers and how dynamic CPU allocation
techniques perform for short-running serverless workloads. We experiment with
Kubernetes as the underlying platform and implement using its vertical pod
autoscaler several dynamic CPU rightsizing techniques. We compare these
techniques using state-of-the-art serverless workloads. Our experiments show
that dynamic CPU allocation for short-running serverless functions is feasible
and can be achieved with lightweight algorithms that offer good performance.
|
Ab initio no-core configuration interaction (NCCI) calculations for the
nuclear many-body problem have traditionally relied upon an antisymmetrized
product (Slater determinant) basis built from harmonic oscillator orbitals. The
accuracy of such calculations is limited by the finite dimensions which are
computationally feasible for the truncated many-body space. We therefore seek
to improve the accuracy obtained for a given basis size by optimizing the
choice of single-particle orbitals. Natural orbitals, which diagonalize the
one-body density matrix, provide a basis which maximizes the occupation of
low-lying orbitals, thus accelerating convergence in a
configuration-interaction basis, while also possibly providing physical insight
into the single-particle structure of the many-body wave function. We describe
the implementation of natural orbitals in the NCCI framework, and examine the
nature of the natural orbitals thus obtained, the properties of the resulting
many-body wave functions, and the convergence of observables. After taking
$^3\mathrm{He}$ as an illustrative testbed, we explore aspects of NCCI
calculations with natural orbitals for the ground state of the $p$-shell
neutron halo nucleus $^6\mathrm{He}$.
|
Logarithmic spirals are conjectured to be optimal escape paths from a half
plane ocean. Assuming this, we find the rate of increase for both min-max and
min-mean interpretations of "optimal". For the one-dimensional analog, which we
call logarithmic coils, our min-mean solution differs from a widely-cited
published account.
|
Context. A large fraction of the interstellar medium can be characterized as
a multiphase medium. The neutral hydrogen gas is bistable with a cold and warm
neutral medium (CNM and WNM respectively) but there is evidence for an
additional phase at intermediate temperatures, a lukewarm neutral medium (LNM)
that is thermally unstable. Aims. We use all sky data from the HI4PI survey to
separate these neutral HI phases with the aim to determine their distribution
and phase fractions in the local interstellar medium. Methods. HI4PI
observations, gridded on an nside = 1024 HEALPix grid, were decomposed into
Gaussian components. From the frequency distribution of the velocity
dispersions we infer three separate linewidth regimes. Accordingly we extract
the HI line emission corresponding to the CNM, LNM, and WNM. We generated
all-sky maps of these phases in the local HI gas with -8 < v_LSR < 8 km/s.
Results. Each of the HI phases shows distinct structures on all scales. The LNM
never exists as a single phase but contributes on average 41% of the HI. The
CNM is prominent only for 22% of the sky, contributes there on average 34% but
locally up to 60% of the HI and is associated with dust at temperatures T_dust
~ 18.6 K. Embedded cold filaments show a clear anti-correlation between CNM and
LNM. Also the smoothly distributed WNM is anti-correlated with the CNM. It
contributes for the rest of the sky 39% with dust associated at temperatures
T_dust ~ 19.4 K. Conclusions. The CNM in filaments exists on small scales. Here
the observed anti-correlation between LNM and CNM implies that both, filaments
and the surrounding more extended LNM, must have a common origin.
|
Bulk high-temperature superconductors (HTS) are capable of generating very
strong magnetic fields while maintaining a relatively compact form factor.
Solenoids constructed using stacks of ring-shaped bulk HTS have been
demonstrated to be capable of nuclear magnetic resonance (NMR) spectroscopy and
magnetic resonance imaging (MRI). However, these stacks were magnetised via
field cooling (FC), which typically requires a secondary superconducting
charging magnet capable of sustaining a high magnetic field for a long period.
A more economical alternative to FC is pulsed field magnetisation, which can be
carried out with a magnet wound from a normal conductor, such as copper. In
this work, we present a technique we have developed for iteratively
homogenising the magnetic field within a stack of ring-shaped bulk HTS by
manipulating the spatial profile of the applied pulsed field.
|
We report the discovery of the galaxy cluster ClJ1226.9+3332 in the Wide
Angle ROSAT Pointed Survey (WARPS). At z=0.888 and L_X=1.1e45 erg/s (0.5-2.0
keV, h_0=0.5) ClJ1226.9+3332 is the most distant X-ray luminous cluster
currently known. The mere existence of this system represents a huge problem
for Omega_0=1 world models.
At the modest (off-axis) resolution of the ROSAT PSPC observation in which
the system was detected, ClJ1226.9+3332 appears relaxed; an off-axis HRI
observation confirms this impression and rules out significant contamination
from point sources. However, in moderately deep optical images (R and I band)
the cluster exhibits signs of substructure in its apparent galaxy distribution.
A first crude estimate of the velocity dispersion of the cluster galaxies based
on six redshifts yields a high value of 1650 km/s, indicative of a very massive
cluster and/or the presence of substructure along the line of sight. While a
more accurate assessment of the dynamical state of this system requires much
better data at both optical and X-ray wavelengths, the high mass of the cluster
has already been unambiguously confirmed by a very strong detection of the
Sunyaev-Zel'dovich effect in its direction (Joy et al. 2001).
Using ClJ1226.9+3332 and ClJ0152.7-1357 (z=0.835), the second-most distant
X-ray luminous cluster currently known and also a WARPS discovery, we obtain a
first estimate of the cluster X-ray luminosity function at 0.8<z<1.4 and
L_X>5e44 erg/s. Using the best currently available data, we find the comoving
space density of very distant, massive clusters to be in excellent agreement
with the value measured locally (z<0.3), and conclude that negative evolution
is not required at these luminosities out to z~1. (truncated)
|
We analyze the dynamics of a magnetic flux quantum (current vortex) trapped
in a current-biased long planar elliptic annular Josephson tunnel junction. The
system is modeled by a perturbed sine-Gordon equation that determines the
spatial and temporal behavior of the phase difference across the tunnel barrier
separating the two superconducting electrodes. In the absence of an external
magnetic field the fluxon dynamics in an elliptic annulus does not differ from
that of a circular annulus where the stationary fluxon speed merely is
determined by the system losses. The interaction between the vortex magnetic
moment and a spatially homogeneous in-plane magnetic field gives rise to a
tunable periodic non-sinusoidal potential which is strongly dependent on the
annulus aspect ratio. We study the escape of the vortex from a well in the
tilted potential when the bias current exceeds the depinning current. The
smallest depinning current as well as the lowest sensitivity of the annulus to
the external field is achieved when the eccentricity is equal to -1. The
presented extensive numerical results are in good agreement with the findings
of the perturbative approach. We also probe the rectifying properties of an
asymmetric potential implemented with an egg-shaped annulus formed by two
semi-elliptic arcs.
|
Decades after the beginning of its FU Orionis-type outburst, V346 Nor
unexpectedly underwent a fading event of $\Delta{}K$ = 4.6 mag around 2010. We
obtained near-infrared observations and re-analysed data from the VISTA/VVV
survey to outline the brightness evolution. In our VLT/NaCO images, we
discovered a halo of scattered light around V346~Nor with a size of about 0.04
arcsec (30 au). The VISTA data outlined a well-defined minimum in the light
curve at late 2010/early 2011, and tentatively revealed a small-amplitude
periodic modulation of 58 days. Our latest data points from 2016 demonstrate
that the source is still brightening but has not reached the 2008 level yet. We
used a simple accretion disk model with varying accretion rate and
line-of-sight extinction to reproduce the observed near-infrared magnitudes and
colors. We found that before 2008, the flux changes of V346 Nor were caused by
a correlated change of extinction and accretion rate, while the minimum around
2010 was mostly due to decreasing accretion. The source reached a maximal
accretion rate of ${\approx}10^{-4} M_{\odot}$ yr$^{-1}$ in 1992. A combination
of accretion and extinction changes was already invoked in the literature to
interpret the flux variations of certain embedded young eruptive stars.
|
We study experimentally a system comprised of linear chains of spin-1/2
nuclei that provides a test-bed for multi-body dynamics and quantum information
processing. This system is a paradigm for a new class of quantum information
devices that can perform particular tasks even without universal control of the
whole quantum system. We investigate the extent of control achievable on the
system with current experimental apparatus and methods to gain information on
the system state, when full tomography is not possible and in any case highly
inefficient.
|
Quasars show a remarkable degree of atomic emission line-broadening, an
observational feature which, in conjunction with a radial distance estimate for
this emission from the nucleus is often used to infer the mass of the central
supermassive black hole. The radius estimate depends on the structure and
kinematics of this so-called Broad-Line Region (BLR), which is often modeled as
a set of discrete emitting clouds. Here, we test an alternative kinematic
disk-wind model of optically thick line emission originating from a
geometrically thin accretion disk under Keplerian rotation around a
supermassive black hole. We use this model to calculate broad emission line
profiles and interferometric phases to compare to GRAVITY data and previously
published cloud modelling results. While we show that such a model can provide
a statistically satisfactory fit to GRAVITY data for quasar 3C 273, we disfavor
it as it requires 3C 273 be observed at high inclination, which observations of
the radio jet orientation do not support.
|
Convolutional neural networks have recently achieved significant
breakthroughs in various image classification tasks. However, they are
computationally expensive,which can make their feasible mplementation on
embedded and low-power devices difficult. In this paper convolutional neural
network binarization is implemented on GPU-based platforms for real-time
inference on resource constrained devices. In binarized networks, all weights
and intermediate computations between layers are quantized to +1 and -1,
allowing multiplications and additions to be replaced with bit-wise operations
between 32-bit words. This representation completely eliminates the need for
floating point multiplications and additions and decreases both the
computational load and the memory footprint compared to a full-precision
network implemented in floating point, making it well-suited for
resource-constrained environments. We compare the performance of our
implementation with an equivalent floating point implementation on one desktop
and two embedded GPU platforms. Our implementation achieves a maximum speed up
of 7. 4X with only 4.4% loss in accuracy compared to a reference
implementation.
|
Thermal fluctuations in non-equilibrium steady states generically lead to
power law decay of correlations for conserved quantities. Embedded bodies which
constrain fluctuations in turn experience fluctuation induced forces. We
compute these forces for the simple case of parallel slabs in a driven
diffusive system. The force falls off with slab separation $d$ as $k_B T/d$ (at
temperature $T$, and in all spatial dimensions), but can be attractive or
repulsive. Unlike the equilibrium Casimir force, the force amplitude is
non-universal and explicitly depends on dynamics. The techniques introduced can
be generalized to study pressure and fluctuation induced forces in a broad
class of non-equilibrium systems.
|
Task-oriented dialogue is often decomposed into three tasks: understanding
user input, deciding actions, and generating a response. While such
decomposition might suggest a dedicated model for each sub-task, we find a
simple, unified approach leads to state-of-the-art performance on the MultiWOZ
dataset. SimpleTOD is a simple approach to task-oriented dialogue that uses a
single, causal language model trained on all sub-tasks recast as a single
sequence prediction problem. This allows SimpleTOD to fully leverage transfer
learning from pre-trained, open domain, causal language models such as GPT-2.
SimpleTOD improves over the prior state-of-the-art in joint goal accuracy for
dialogue state tracking, and our analysis reveals robustness to noisy
annotations in this setting. SimpleTOD also improves the main metrics used to
evaluate action decisions and response generation in an end-to-end setting:
inform rate by 8.1 points, success rate by 9.7 points, and combined score by
7.2 points.
|
Mirrors are ubiquitous in optics and are used to control the propagation of
optical signals in space. Here we propose and demonstrate frequency domain
mirrors that provide reflections of the optical energy in a frequency synthetic
dimension, using electro-optic modulation. First, we theoretically explore the
concept of frequency mirrors with the investigation of propagation loss, and
reflectivity in the frequency domain. Next, we explore the mirror formed
through polarization mode-splitting in a thin-film lithium niobate
micro-resonator. By exciting the Bloch waves of the synthetic frequency crystal
with different wave vectors, we show various states formed by the interference
between forward propagating and reflected waves. Finally, we expand on this
idea, and generate tunable frequency mirrors as well as demonstrate trapped
states formed by these mirrors using coupled lithium niobate micro-resonators.
The ability to control the flow of light in the frequency domain could enable a
wide range of applications, including the study of random walks, boson
sampling, frequency comb sources, optical computation, and topological
photonics. Furthermore, demonstration of optical elements such as cavities,
lasers, and photonic crystals in the frequency domain, may be possible.
|
Many modern and proposed future particle accelerators rely on superconducting
radio frequency cavities made of bulk niobium as primary particle accelerating
structures. Such cavities suffer from the anomalous field dependence of their
quality factors Q0. High field degradation - so-called high field Q-slope - is
yet unexplained even though an empirical cure is known. Here we propose a
mechanism based on the presence of proximity-coupled niobium hydrides, which
can explain this effect. Furthermore, the same mechanism can be present in any
surface-sensitive experiments or superconducting devices involving niobium.
|
Cross-domain sentiment classification (CDSC) is an importance task in domain
adaptation and sentiment classification. Due to the domain discrepancy, a
sentiment classifier trained on source domain data may not works well on target
domain data. In recent years, many researchers have used deep neural network
models for cross-domain sentiment classification task, many of which use
Gradient Reversal Layer (GRL) to design an adversarial network structure to
train a domain-shared sentiment classifier. Different from those methods, we
proposed Hierarchical Attention Generative Adversarial Networks (HAGAN) which
alternately trains a generator and a discriminator in order to produce a
document representation which is sentiment-distinguishable but
domain-indistinguishable. Besides, the HAGAN model applies Bidirectional Gated
Recurrent Unit (Bi-GRU) to encode the contextual information of a word and a
sentence into the document representation. In addition, the HAGAN model use
hierarchical attention mechanism to optimize the document representation and
automatically capture the pivots and non-pivots. The experiments on Amazon
review dataset show the effectiveness of HAGAN.
|
This study explores the problem solving capabilities of ChatGPT and its
prospective applications in standardized test preparation, focusing on the GRE
quantitative exam. Prior research has shown great potential for the utilization
of ChatGPT for academic purposes in revolutionizing the approach to studying
across various disciplines. We investigate how ChatGPT performs across various
question types in the GRE quantitative domain, and how modifying question
prompts impacts its accuracy. More specifically this study addressed two
research questions: 1. How does ChatGPT perform in answering GRE-based
quantitative questions across various content areas? 2. How does the accuracy
of ChatGPT vary with modifying the question prompts? The dataset consisting of
100 randomly selected GRE quantitative questions was collected from the ETS
official guide to GRE test preparation. We used quantitative evaluation to
answer our first research question, and t-test to examine the statistical
association between prompt modification and ChatGPT's accuracy. Results show a
statistical improvement in the ChatGPT's accuracy after applying instruction
priming and contextual prompts to the original questions. ChatGPT showed 84%
accuracy with the modified prompts compared to 69% with the original data. The
study discusses the areas where ChatGPT struggled with certain questions and
how modifications can be helpful for preparing for standardized tests like GRE
and provides future directions for prompt modifications.
|
Future developments in deep learning applications requiring large datasets
will be limited by power and speed limitations of silicon based Von-Neumann
computing architectures. Optical architectures provide a low power and high
speed hardware alternative. Recent publications have suggested promising
implementations of optical neural networks (ONNs), showing huge orders of
magnitude efficiency and speed gains over current state of the art hardware
alternatives. In this work, the transmission of the Fabry-Perot Interferometer
(FPI) is proposed as a low power, low footprint activation function unit.
Numerical simulations of optical CNNs using the FPI based activation functions
show accuracies of 98% on the MNIST dataset. An investigation of possible
physical implementation of the network shows that an ONN based on current
tunable FPIs could be slowed by actuation delays, but rapidly developing
optical hardware fabrication techniques could make an integrated approach using
the proposed FPI setups a powerful solution for previously inaccessible deep
learning applications.
|
Unordered data Petri nets (UDPN) are an extension of classical Petri nets
with tokens that carry data from an infinite domain and where transitions may
check equality and disequality of tokens. UDPN are well-structured, so the
coverability and termination problems are decidable, but with higher complexity
than for Petri nets. On the other hand, the problem of reachability for UDPN is
surprisingly complex, and its decidability status remains open. In this paper,
we consider the continuous reachability problem for UDPN, which can be seen as
an over-approximation of the reachability problem. Our main result is a
characterization of continuous reachability for UDPN and polynomial time
algorithm for solving it. This is a consequence of a combinatorial argument,
which shows that if continuous reachability holds then there exists a run using
only polynomially many data values.
|
The field of cybersecurity is evolving fast. Experts need to be informed
about past, current and - in the best case - upcoming threats, because attacks
are becoming more advanced, targets bigger and systems more complex. As this
cannot be addressed manually, cybersecurity experts need to rely on machine
learning techniques. In the texutual domain, pre-trained language models like
BERT have shown to be helpful, by providing a good baseline for further
fine-tuning. However, due to the domain-knowledge and many technical terms in
cybersecurity general language models might miss the gist of textual
information, hence doing more harm than good. For this reason, we create a
high-quality dataset and present a language model specifically tailored to the
cybersecurity domain, which can serve as a basic building block for
cybersecurity systems that deal with natural language. The model is compared
with other models based on 15 different domain-dependent extrinsic and
intrinsic tasks as well as general tasks from the SuperGLUE benchmark. On the
one hand, the results of the intrinsic tasks show that our model improves the
internal representation space of words compared to the other models. On the
other hand, the extrinsic, domain-dependent tasks, consisting of sequence
tagging and classification, show that the model is best in specific application
scenarios, in contrast to the others. Furthermore, we show that our approach
against catastrophic forgetting works, as the model is able to retrieve the
previously trained domain-independent knowledge. The used dataset and trained
model are made publicly available
|
Subsets and Splits