title
stringlengths 7
239
| abstract
stringlengths 7
2.76k
| cs
int64 0
1
| phy
int64 0
1
| math
int64 0
1
| stat
int64 0
1
| quantitative biology
int64 0
1
| quantitative finance
int64 0
1
|
|---|---|---|---|---|---|---|---|
Memory effects, transient growth, and wave breakup in a model of paced atrium | The mechanisms underlying cardiac fibrillation have been investigated for
over a century, but we are still finding surprising results that change our
view of this phenomenon. The present study focuses on the transition from
normal rhythm to atrial fibrillation associated with a gradual increase in the
pacing rate. While some of our findings are consistent with existing
experimental, numerical, and theoretical studies of this problem, one result
appears to contradict the accepted picture. Specifically we show that, in a
two-dimensional model of paced homogeneous atrial tissue, transition from
discordant alternans to conduction block, wave breakup, reentry, and spiral
wave chaos is associated with transient growth of finite amplitude disturbances
rather than a conventional instability. It is mathematically very similar to
subcritical, or bypass, transition from laminar fluid flow to turbulence, which
allows many of the tools developed in the context of fluid turbulence to be
used for improving our understanding of cardiac arrhythmias.
| 0 | 1 | 0 | 0 | 0 | 0 |
An Analysis of the Value of Information when Exploring Stochastic, Discrete Multi-Armed Bandits | In this paper, we propose an information-theoretic exploration strategy for
stochastic, discrete multi-armed bandits that achieves optimal regret. Our
strategy is based on the value of information criterion. This criterion
measures the trade-off between policy information and obtainable rewards. High
amounts of policy information are associated with exploration-dominant searches
of the space and yield high rewards. Low amounts of policy information favor
the exploitation of existing knowledge. Information, in this criterion, is
quantified by a parameter that can be varied during search. We demonstrate that
a simulated-annealing-like update of this parameter, with a sufficiently fast
cooling schedule, leads to an optimal regret that is logarithmic with respect
to the number of episodes.
| 1 | 0 | 0 | 1 | 0 | 0 |
Probabilistic Generative Adversarial Networks | We introduce the Probabilistic Generative Adversarial Network (PGAN), a new
GAN variant based on a new kind of objective function. The central idea is to
integrate a probabilistic model (a Gaussian Mixture Model, in our case) into
the GAN framework which supports a new kind of loss function (based on
likelihood rather than classification loss), and at the same time gives a
meaningful measure of the quality of the outputs generated by the network.
Experiments with MNIST show that the model learns to generate realistic images,
and at the same time computes likelihoods that are correlated with the quality
of the generated images. We show that PGAN is better able to cope with
instability problems that are usually observed in the GAN training procedure.
We investigate this from three aspects: the probability landscape of the
discriminator, gradients of the generator, and the perfect discriminator
problem.
| 1 | 0 | 0 | 1 | 0 | 0 |
Model comparison for Gibbs random fields using noisy reversible jump Markov chain Monte Carlo | The reversible jump Markov chain Monte Carlo (RJMCMC) method offers an
across-model simulation approach for Bayesian estimation and model comparison,
by exploring the sampling space that consists of several models of possibly
varying dimensions. A naive implementation of RJMCMC to models like Gibbs
random fields suffers from computational difficulties: the posterior
distribution for each model is termed doubly-intractable since computation of
the likelihood function is rarely available. Consequently, it is simply
impossible to simulate a transition of the Markov chain in the presence of
likelihood intractability. A variant of RJMCMC is presented, called noisy
RJMCMC, where the underlying transition kernel is replaced with an
approximation based on unbiased estimators. Based on previous theoretical
developments, convergence guarantees for the noisy RJMCMC algorithm are
provided. The experiments show that the noisy RJMCMC algorithm can be much more
efficient than other exact methods, provided that an estimator with controlled
Monte Carlo variance is used, a fact which is in agreement with the theoretical
analysis.
| 0 | 0 | 0 | 1 | 0 | 0 |
Functorial compactification of linear spaces | We define compactifications of vector spaces which are functorial with
respect to certain linear maps. These "many-body" compactifications are
manifolds with corners, and the linear maps lift to b-maps in the sense of
Melrose. We derive a simple criterion under which the lifted maps are in fact
b-fibrations, and identify how these restrict to boundary hypersurfaces. This
theory is an application of a general result on the iterated blow-up of cleanly
intersecting submanifolds which extends related results in the literature.
| 0 | 0 | 1 | 0 | 0 | 0 |
Almost complex structures on connected sums of complex projective spaces | We show that the m-fold connected sum $m\#\mathbb{C}\mathbb{P}^{2n}$ admits
an almost complex structure if and only if m is odd.
| 0 | 0 | 1 | 0 | 0 | 0 |
Raman Scattering by a Two-Dimensional Fermi Liquid with Spin-Orbit Coupling | We present a microscopic theory of Raman scattering by a two-dimensional
Fermi liquid (FL) with Rashba and Dresselhaus types of spin-orbit coupling, and
subject to an in-plane magnetic field (B). In the long-wavelength limit, the
Raman spectrum probes the collective modes of such a FL: the chiral spin waves.
The characteristic features of these modes are a linear-in-q term in the
dispersion and the dependence of the mode frequency on the directions of both q
and B. All of these features have been observed in recent Raman experiments on
CdTe quantum wells.
| 0 | 1 | 0 | 0 | 0 | 0 |
Nearly Optimal Robust Subspace Tracking | In this work, we study the robust subspace tracking (RST) problem and obtain
one of the first two provable guarantees for it. The goal of RST is to track
sequentially arriving data vectors that lie in a slowly changing
low-dimensional subspace, while being robust to corruption by additive sparse
outliers. It can also be interpreted as a dynamic (time-varying) extension of
robust PCA (RPCA), with the minor difference that RST also requires a short
tracking delay. We develop a recursive projected compressive sensing algorithm
that we call Nearly Optimal RST via ReProCS (ReProCS-NORST) because its
tracking delay is nearly optimal. We prove that NORST solves both the RST and
the dynamic RPCA problems under weakened standard RPCA assumptions, two simple
extra assumptions (slow subspace change and most outlier magnitudes lower
bounded), and a few minor assumptions.
Our guarantee shows that NORST enjoys a near optimal tracking delay of $O(r
\log n \log(1/\epsilon))$. Its required delay between subspace change times is
the same, and its memory complexity is $n$ times this value. Thus both these
are also nearly optimal. Here $n$ is the ambient space dimension, $r$ is the
subspaces' dimension, and $\epsilon$ is the tracking accuracy. NORST also has
the best outlier tolerance compared with all previous RPCA or RST methods, both
theoretically and empirically (including for real videos), without requiring
any model on how the outlier support is generated. This is possible because of
the extra assumptions it uses.
| 0 | 0 | 0 | 1 | 0 | 0 |
The Authority of "Fair" in Machine Learning | In this paper, we argue for the adoption of a normative definition of
fairness within the machine learning community. After characterizing this
definition, we review the current literature of Fair ML in light of its
implications. We end by suggesting ways to incorporate a broader community and
generate further debate around how to decide what is fair in ML.
| 1 | 0 | 0 | 0 | 0 | 0 |
The Social Bow Tie | Understanding tie strength in social networks, and the factors that influence
it, have received much attention in a myriad of disciplines for decades.
Several models incorporating indicators of tie strength have been proposed and
used to quantify relationships in social networks, and a standard set of
structural network metrics have been applied to predominantly online social
media sites to predict tie strength. Here, we introduce the concept of the
"social bow tie" framework, a small subgraph of the network that consists of a
collection of nodes and ties that surround a tie of interest, forming a
topological structure that resembles a bow tie. We also define several
intuitive and interpretable metrics that quantify properties of the bow tie. We
use random forests and regression models to predict categorical and continuous
measures of tie strength from different properties of the bow tie, including
nodal attributes. We also investigate what aspects of the bow tie are most
predictive of tie strength in two distinct social networks: a collection of 75
rural villages in India and a nationwide call network of European mobile phone
users. Our results indicate several of the bow tie metrics are highly
predictive of tie strength, and we find the more the social circles of two
individuals overlap, the stronger their tie, consistent with previous findings.
However, we also find that the more tightly-knit their non-overlapping social
circles, the weaker the tie. This new finding complements our current
understanding of what drives the strength of ties in social networks.
| 0 | 0 | 0 | 1 | 0 | 0 |
Response Regimes in Equivalent Mechanical Model of Moderately Nonlinear Liquid Sloshing | The paper considers non-stationary responses in reduced-order model of
partially liquid-filled tank under external forcing. The model involves one
common degree of freedom for the tank and the non-sloshing portion of the
liquid, and the other one -- for the sloshing portion of the liquid. The
coupling between these degrees of freedom is nonlinear, with the lowest-order
potential dictated by symmetry considerations. Since the mass of the sloshing
liquid in realistic conditions does not exceed 10% of the total mass of the
system, the reduced-order model turns to be formally equivalent to well-studied
oscillatory systems with nonlinear energy sinks (NES). Exploiting this analogy,
and applying the methodology known from the studies of the systems with the
NES, we predict a multitude of possible non-stationary responses in the
considered model. These responses conform, at least on the qualitative level,
to the responses observed in experimental sloshing settings, multi-modal
theoretical models and full-scale numeric simulations.
| 0 | 1 | 0 | 0 | 0 | 0 |
Opinion evolution in time-varying social influence networks with prejudiced agents | Investigation of social influence dynamics requires mathematical models that
are "simple" enough to admit rigorous analysis, and yet sufficiently "rich" to
capture salient features of social groups. Thus, the mechanism of iterative
opinion pooling from (DeGroot, 1974), which can explain the generation of
consensus, was elaborated in (Friedkin and Johnsen, 1999) to take into account
individuals' ongoing attachments to their initial opinions, or prejudices. The
"anchorage" of individuals to their prejudices may disable reaching consensus
and cause disagreement in a social influence network. Further elaboration of
this model may be achieved by relaxing its restrictive assumption of a
time-invariant influence network. During opinion dynamics on an issue, arcs of
interpersonal influence may be added or subtracted from the network, and the
influence weights assigned by an individual to his/her neighbors may alter. In
this paper, we establish new important properties of the (Friedkin and Johnsen,
1999) opinion formation model, and also examine its extension to time-varying
social influence networks.
| 1 | 1 | 1 | 0 | 0 | 0 |
k*-Nearest Neighbors: From Global to Local | The weighted k-nearest neighbors algorithm is one of the most fundamental
non-parametric methods in pattern recognition and machine learning. The
question of setting the optimal number of neighbors as well as the optimal
weights has received much attention throughout the years, nevertheless this
problem seems to have remained unsettled. In this paper we offer a simple
approach to locally weighted regression/classification, where we make the
bias-variance tradeoff explicit. Our formulation enables us to phrase a notion
of optimal weights, and to efficiently find these weights as well as the
optimal number of neighbors efficiently and adaptively, for each data point
whose value we wish to estimate. The applicability of our approach is
demonstrated on several datasets, showing superior performance over standard
locally weighted methods.
| 1 | 0 | 0 | 1 | 0 | 0 |
Network Slicing for Ultra-Reliable Low Latency Communication in Industry 4.0 Scenarios | An important novelty of 5G is its role in transforming the industrial
production into Industry 4.0. Specifically, Ultra-Reliable Low Latency
Communications (URLLC) will, in many cases, enable replacement of cables with
wireless connections and bring freedom in designing and operating
interconnected machines, robots, and devices. However, not all industrial links
will be of URLLC type; e.g. some applications will require high data rates.
Furthermore, these industrial networks will be highly heterogeneous, featuring
various communication technologies. We consider network slicing as a mechanism
to handle the diverse set of requirements to the network. We present methods
for slicing deterministic and packet-switched industrial communication
protocols at an abstraction level that is decoupled from the specific
implementation of the underlying technologies. Finally, we show how network
calculus can be used to assess the end-to-end properties of the network slices.
| 1 | 0 | 0 | 0 | 0 | 0 |
Markov Decision Processes with Continuous Side Information | We consider a reinforcement learning (RL) setting in which the agent
interacts with a sequence of episodic MDPs. At the start of each episode the
agent has access to some side-information or context that determines the
dynamics of the MDP for that episode. Our setting is motivated by applications
in healthcare where baseline measurements of a patient at the start of a
treatment episode form the context that may provide information about how the
patient might respond to treatment decisions. We propose algorithms for
learning in such Contextual Markov Decision Processes (CMDPs) under an
assumption that the unobserved MDP parameters vary smoothly with the observed
context. We also give lower and upper PAC bounds under the smoothness
assumption. Because our lower bound has an exponential dependence on the
dimension, we consider a tractable linear setting where the context is used to
create linear combinations of a finite set of MDPs. For the linear setting, we
give a PAC learning algorithm based on KWIK learning techniques.
| 1 | 0 | 0 | 1 | 0 | 0 |
Nonlinear stage of Benjamin-Feir instability in forced/damped deep water waves | We study a three-wave truncation of a recently proposed damped/forced
high-order nonlinear Schrödinger equation for deep-water gravity waves under
the effect of wind and viscosity. The evolution of the norm (wave-action) and
spectral mean of the full model are well captured by the reduced dynamics.
Three regimes are found for the wind-viscosity balance: we classify them
according to the attractor in the phase-plane of the truncated system and to
the shift of the spectral mean. A downshift can coexist with both net forcing
and damping, i.e., attraction to period-1 or period-2 solutions. Upshift is
associated with stronger winds, i.e., to a net forcing where the attractor is
always a period-1 solution. The applicability of our classification to
experiments in long wave-tanks is verified.
| 0 | 1 | 0 | 0 | 0 | 0 |
Computational Thinking in Patch | With the future likely to see even more pervasive computation, computational
thinking (problem-solving skills incorporating computing knowledge) is now
being recognized as a fundamental skill needed by all students. Computational
thinking is conceptualizing as opposed to programming, promotes natural human
thinking style than algorithmic reasoning, complements and combines
mathematical and engineering thinking, and it emphasizes ideas, not artifacts.
In this paper, we outline a new visual language, called Patch, using which
students are able to express their solutions to eScience computational problems
in abstract visual tools. Patch is closer to high level procedural languages
such as C++ or Java than Scratch or Snap! but similar to them in ease of use
and combines simplicity and expressive power in one single platform.
| 1 | 0 | 0 | 0 | 0 | 0 |
Skoda's Ideal Generation from Vanishing Theorem for Semipositive Nakano Curvature and Cauchy-Schwarz Inequality for Tensors | Skoda's 1972 result on ideal generation is a crucial ingredient in the
analytic approach to the finite generation of the canonical ring and the
abundance conjecture. Special analytic techniques developed by Skoda, other
than applications of the usual vanishing theorems and L2 estimates for the
d-bar equation, are required for its proof. This note (which is part of a
lecture given in the 60th birthday conference for Lawrence Ein) gives a
simpler, more straightforward proof of Skoda's result, which makes it a natural
consequence of the standard techniques in vanishing theorems and solving d-bar
equation with L2 estimates. The proof involves the following three ingredients:
(i) one particular Cauchy-Schwarz inequality for tensors with a special factor
which accounts for the exponent of the denominator in the formulation of the
integral condition for Skoda's ideal generation, (ii) the nonnegativity of
Nakano curvature of the induced metric of a special co-rank-1 subbundle of a
trivial vector bundle twisted by a special scalar weight function, and (iii)
the vanishing theorem and solvability of d-bar equation with L2 estimates for
vector bundles of nonnegative Nakano curvature on a strictly pseudoconvex
domain. Our proof gives readily other similar results on ideal generation.
| 0 | 0 | 1 | 0 | 0 | 0 |
Computing Nonvacuous Generalization Bounds for Deep (Stochastic) Neural Networks with Many More Parameters than Training Data | One of the defining properties of deep learning is that models are chosen to
have many more parameters than available training data. In light of this
capacity for overfitting, it is remarkable that simple algorithms like SGD
reliably return solutions with low test error. One roadblock to explaining
these phenomena in terms of implicit regularization, structural properties of
the solution, and/or easiness of the data is that many learning bounds are
quantitatively vacuous when applied to networks learned by SGD in this "deep
learning" regime. Logically, in order to explain generalization, we need
nonvacuous bounds. We return to an idea by Langford and Caruana (2001), who
used PAC-Bayes bounds to compute nonvacuous numerical bounds on generalization
error for stochastic two-layer two-hidden-unit neural networks via a
sensitivity analysis. By optimizing the PAC-Bayes bound directly, we are able
to extend their approach and obtain nonvacuous generalization bounds for deep
stochastic neural network classifiers with millions of parameters trained on
only tens of thousands of examples. We connect our findings to recent and old
work on flat minima and MDL-based explanations of generalization.
| 1 | 0 | 0 | 0 | 0 | 0 |
Method for Computationally Efficient Design of Dielectric Laser Accelerators | Dielectric microstructures have generated much interest in recent years as a
means of accelerating charged particles when powered by solid state lasers. The
acceleration gradient (or particle energy gain per unit length) is an important
figure of merit. To design structures with high acceleration gradients, we
explore the adjoint variable method, a highly efficient technique used to
compute the sensitivity of an objective with respect to a large number of
parameters. With this formalism, the sensitivity of the acceleration gradient
of a dielectric structure with respect to its entire spatial permittivity
distribution is calculated by the use of only two full-field electromagnetic
simulations, the original and adjoint. The adjoint simulation corresponds
physically to the reciprocal situation of a point charge moving through the
accelerator gap and radiating. Using this formalism, we perform numerical
optimizations aimed at maximizing acceleration gradients, which generate
fabricable structures of greatly improved performance in comparison to
previously examined geometries.
| 0 | 1 | 0 | 0 | 0 | 0 |
Computing and Using Minimal Polynomials | Given a zero-dimensional ideal I in a polynomial ring, many computations
start by finding univariate polynomials in I. Searching for a univariate
polynomial in I is a particular case of considering the minimal polynomial of
an element in P/I. It is well known that minimal polynomials may be computed
via elimination, therefore this is considered to be a "resolved problem". But
being the key of so many computations, it is worth investigating its meaning,
its optimization, its applications.
| 1 | 0 | 1 | 0 | 0 | 0 |
DiVM: Model Checking with LLVM and Graph Memory | In this paper, we introduce the concept of a virtual machine with
graph-organised memory as a versatile backend for both explicit-state and
abstraction-driven verification of software. Our virtual machine uses the LLVM
IR as its instruction set, enriched with a small set of hypercalls. We show
that the provided hypercalls are sufficient to implement a small operating
system, which can then be linked with applications to provide a
POSIX-compatible verification environment. Finally, we demonstrate the
viability of the approach through a comparison with a more
traditionally-designed LLVM model checker.
| 1 | 0 | 0 | 0 | 0 | 0 |
Uhlenbeck's decomposition in Sobolev and Morrey-Sobolev spaces | We present a self-contained proof of Uhlenbeck's decomposition theorem for
$\Omega\in L^p(\mathbb{B}^n,so(m)\otimes\Lambda^1\mathbb{R}^n)$ for $p\in
(1,n)$ with Sobolev type estimates in the case $p \in[n/2,n)$ and
Morrey-Sobolev type estimates in the case $p\in (1,n/2)$. We also prove an
analogous theorem in the case when $\Omega\in L^p( \mathbb{B}^n, TCO_{+}(m)
\otimes \Lambda^1\mathbb{R}^n)$, which corresponds to Uhlenbeck's theorem with
conformal gauge group.
| 0 | 0 | 1 | 0 | 0 | 0 |
Making Asynchronous Distributed Computations Robust to Noise | We consider the problem of making distributed computations robust to noise,
in particular to worst-case (adversarial) corruptions of messages. We give a
general distributed interactive coding scheme which simulates any asynchronous
distributed protocol while tolerating an optimal corruption of a $\Theta(1/n)$
fraction of all messages while incurring a moderate blowup of $O(n\log^2 n)$ in
the communication complexity.
Our result is the first fully distributed interactive coding scheme in which
the topology of the communication network is not known in advance. Prior work
required either a coordinating node to be connected to all other nodes in the
network or assumed a synchronous network in which all nodes already know the
complete topology of the network.
| 1 | 0 | 0 | 0 | 0 | 0 |
Justifications in Constraint Handling Rules for Logical Retraction in Dynamic Algorithms | We present a straightforward source-to-source transformation that introduces
justifications for user-defined constraints into the CHR programming language.
Then a scheme of two rules suffices to allow for logical retraction (deletion,
removal) of constraints during computation. Without the need to recompute from
scratch, these rules remove not only the constraint but also undo all
consequences of the rule applications that involved the constraint. We prove a
confluence result concerning the rule scheme and show its correctness. When
algorithms are written in CHR, constraints represent both data and operations.
CHR is already incremental by nature, i.e. constraints can be added at runtime.
Logical retraction adds decrementality. Hence any algorithm written in CHR with
justifications will become fully dynamic. Operations can be undone and data can
be removed at any point in the computation without compromising the correctness
of the result. We present two classical examples of dynamic algorithms, written
in our prototype implementation of CHR with justifications that is available
online: maintaining the minimum of a changing set of numbers and shortest paths
in a graph whose edges change.
| 1 | 0 | 0 | 0 | 0 | 0 |
Bose-Hubbard lattice as a controllable environment for open quantum systems | We investigate the open dynamics of an atomic impurity embedded in a
one-dimensional Bose-Hubbard lattice. We derive the reduced evolution equation
for the impurity and show that the Bose-Hubbard lattice behaves as a tunable
engineered environment allowing to simulate both Markovian and non-Markovian
dynamics in a controlled and experimentally realisable way. We demonstrate that
the presence or absence of memory effects is a signature of the nature of the
excitations induced by the impurity, being delocalized or localized in the two
limiting cases of superfluid and Mott insulator, respectively. Furthermore, our
findings show how the excitations supported in the two phases can be
characterized as information carriers.
| 0 | 1 | 0 | 0 | 0 | 0 |
Semi-decidable equivalence relations obtained by composition and lattice join of decidable equivalence relations | Composition and lattice join (transitive closure of a union) of equivalence
relations are operations taking pairs of decidable equivalence relations to
relations that are semi-decidable, but not necessarily decidable. This article
addresses the question, is every semi-decidable equivalence relation obtainable
in those ways from a pair of decidable equivalence relations? It is shown that
every semi-decidable equivalence relation, of which every equivalence class is
infinite, is obtainable as both a composition and a lattice join of decidable
equivalence relations having infinite equivalence classes. An example is
constructed of a semi-decidable, but not decidable, equivalence relation having
finite equivalence classes that can be obtained from decidable equivalence
relations, both by composition and also by lattice join. Another example is
constructed, in which such a relation cannot be obtained from decidable
equivalence relations in either of the two ways.
| 0 | 0 | 1 | 0 | 0 | 0 |
ClipAudit: A Simple Risk-Limiting Post-Election Audit | We propose a simple risk-limiting audit for elections, ClipAudit. To
determine whether candidate A (the reported winner) actually beat candidate B
in a plurality election, ClipAudit draws ballots at random, without
replacement, until either all cast ballots have been drawn, or until \[ a - b
\ge \beta \sqrt{a+b}
\] where $a$ is the number of ballots in the sample for the reported winner
A, and $b$ is the number of ballots in the sample for opponent B, and where
$\beta$ is a constant determined a priori as a function of the number $n$ of
ballots cast and the risk-limit $\alpha$. ClipAudit doesn't depend on the
unofficial margin (as does Bravo). We show how to extend ClipAudit to contests
with multiple winners or losers, or to multiple contests.
| 1 | 0 | 0 | 1 | 0 | 0 |
LCA(2), Weil index, and product formula | In this paper we study the category LCA(2) of certain non-locally compact
abelian topological groups, and extend the notion of Weil index. As
applications we deduce some product formulas for curves over local fields and
arithmetic surfaces.
| 0 | 0 | 1 | 0 | 0 | 0 |
A Dichotomy for Sampling Barrier-Crossing Events of Random Walks with Regularly Varying Tails | We study how to sample paths of a random walk up to the first time it crosses
a fixed barrier, in the setting where the step sizes are iid with negative mean
and have a regularly varying right tail. We introduce a desirable property for
a change of measure to be suitable for exact simulation. We study whether the
change of measure of Blanchet and Glynn (2008) satisfies this property and show
that it does so if and only if the tail index $\alpha$ of the right tail lies
in the interval $(1, \, 3/2)$.
| 0 | 0 | 1 | 1 | 0 | 0 |
Crawling migration under chemical signalling: a stochastic particle model | Cell migration is a fundamental process involved in physiological phenomena
such as the immune response and morphogenesis, but also in pathological
processes, such as the development of tumor metastasis. These functions are
effectively ensured because cells are active systems that adapt to their
environment. In this work, we consider a migrating cell as an active particle,
where its intracellular activity is responsible for motion. Such system was
already modeled in a previous model where the protrusion activity of the cell
was described by a stochastic Markovian jump process. The model was proven able
to capture the diversity in observed trajectories. Here, we add a description
of the effect of an external chemical attractive signal on the protrusion
dynamics, that may vary in time. We show that the resulting stochastic model is
a well-posed non-homogeneous Markovian process, and provide cell trajectories
in different settings, illustrating the effects of the signal on long-term
trajectories.
| 0 | 0 | 0 | 0 | 1 | 0 |
Towards a Deeper Understanding of Adversarial Losses | Recent work has proposed various adversarial losses for training generative
adversarial networks. Yet, it remains unclear what certain types of functions
are valid adversarial loss functions, and how these loss functions perform
against one another. In this paper, we aim to gain a deeper understanding of
adversarial losses by decoupling the effects of their component functions and
regularization terms. We first derive some necessary and sufficient conditions
of the component functions such that the adversarial loss is a divergence-like
measure between the data and the model distributions. In order to
systematically compare different adversarial losses, we then propose DANTest, a
new, simple framework based on discriminative adversarial networks. With this
framework, we evaluate an extensive set of adversarial losses by combining
different component functions and regularization approaches. This study leads
to some new insights into the adversarial losses. For reproducibility, all
source code is available at this https URL .
| 1 | 0 | 0 | 1 | 0 | 0 |
Transit Visibility Zones of the Solar System Planets | The detection of thousands of extrasolar planets by the transit method
naturally raises the question of whether potential extrasolar observers could
detect the transits of the Solar System planets. We present a comprehensive
analysis of the regions in the sky from where transit events of the Solar
System planets can be detected. We specify how many different Solar System
planets can be observed from any given point in the sky, and find the maximum
number to be three. We report the probabilities of a randomly positioned
external observer to be able to observe single and multiple Solar System planet
transits; specifically, we find a probability of 2.518% to be able to observe
at least one transiting planet, 0.229% for at least two transiting planets, and
0.027% for three transiting planets. We identify 68 known exoplanets that have
a favourable geometric perspective to allow transit detections in the Solar
System and we show how the ongoing K2 mission will extend this list. We use
occurrence rates of exoplanets to estimate that there are $3.2\pm1.2$ and
$6.6^{+1.3}_{-0.8}$ temperate Earth-sized planets orbiting GK and M dwarf stars
brighter than $V=13$ and $V=16$ respectively, that are located in the Earth's
transit zone.
| 0 | 1 | 0 | 0 | 0 | 0 |
Nearest-neighbour Markov point processes on graphs with Euclidean edges | We define nearest-neighbour point processes on graphs with Euclidean edges
and linear networks. They can be seen as the analogues of renewal processes on
the real line. We show that the Delaunay neighbourhood relation on a tree
satisfies the Baddeley--M{\o}ller consistency conditions and provide a
characterisation of Markov functions with respect to this relation. We show
that a modified relation defined in terms of the local geometry of the graph
satisfies the consistency conditions for all graphs with Euclidean edges.
| 0 | 0 | 1 | 1 | 0 | 0 |
A Hierarchical Bayesian Linear Regression Model with Local Features for Stochastic Dynamics Approximation | One of the challenges in model-based control of stochastic dynamical systems
is that the state transition dynamics are involved, and it is not easy or
efficient to make good-quality predictions of the states. Moreover, there are
not many representational models for the majority of autonomous systems, as it
is not easy to build a compact model that captures the entire dynamical
subtleties and uncertainties. In this work, we present a hierarchical Bayesian
linear regression model with local features to learn the dynamics of a
micro-robotic system as well as two simpler examples, consisting of a
stochastic mass-spring damper and a stochastic double inverted pendulum on a
cart. The model is hierarchical since we assume non-stationary priors for the
model parameters. These non-stationary priors make the model more flexible by
imposing priors on the priors of the model. To solve the maximum likelihood
(ML) problem for this hierarchical model, we use the variational expectation
maximization (EM) algorithm, and enhance the procedure by introducing hidden
target variables. The algorithm yields parsimonious model structures, and
consistently provides fast and accurate predictions for all our examples
involving large training and test sets. This demonstrates the effectiveness of
the method in learning stochastic dynamics, which makes it suitable for future
use in a paradigm, such as model-based reinforcement learning, to compute
optimal control policies in real time.
| 0 | 0 | 0 | 1 | 0 | 0 |
Multitask Learning and Benchmarking with Clinical Time Series Data | Health care is one of the most exciting frontiers in data mining and machine
learning. Successful adoption of electronic health records (EHRs) created an
explosion in digital clinical data available for analysis, but progress in
machine learning for healthcare research has been difficult to measure because
of the absence of publicly available benchmark data sets. To address this
problem, we propose four clinical prediction benchmarks using data derived from
the publicly available Medical Information Mart for Intensive Care (MIMIC-III)
database. These tasks cover a range of clinical problems including modeling
risk of mortality, forecasting length of stay, detecting physiologic decline,
and phenotype classification. We propose strong linear and neural baselines for
all four tasks and evaluate the effect of deep supervision, multitask training
and data-specific architectural modifications on the performance of neural
models.
| 1 | 0 | 0 | 1 | 0 | 0 |
Essentially Finite Vector Bundles on Normal Pseudo-proper Algebraic Stacks | Let $X$ be a normal, connected and projective variety over an algebraically
closed field $k$. It is known that a vector bundle $V$ on $X$ is essentially
finite if and only if it is trivialized by a proper surjective morphism $f:Y\to
X$. In this paper we introduce a different approach to this problem which
allows to extend the results to normal, connected and strongly pseudo-proper
algebraic stack of finite type over an arbitrary field $k$.
| 0 | 0 | 1 | 0 | 0 | 0 |
Fluid flow across a wavy channel brought in contact | A pressure driven flow in contact interface between elastic solids with wavy
surfaces is studied. We consider a strong coupling between the solid and the
fluid problems, which is relevant when the fluid pressure is comparable with
the contact pressure. An approximate analytical solution is obtained for this
coupled problem. A finite-element monolithically coupled framework is used to
solve the problem numerically. A good agreement is obtained between the two
solutions within the region of the validity of the analytical one. A power-law
interface transmissivity decay is observed near the percolation. Finally, we
showed that the external pressure needed to seal the channel is an affine
function of the inlet pressure and does not depend on the outlet pressure.
| 0 | 1 | 0 | 0 | 0 | 0 |
Species tree estimation using ASTRAL: how many genes are enough? | Species tree reconstruction from genomic data is increasingly performed using
methods that account for sources of gene tree discordance such as incomplete
lineage sorting. One popular method for reconstructing species trees from
unrooted gene tree topologies is ASTRAL. In this paper, we derive theoretical
sample complexity results for the number of genes required by ASTRAL to
guarantee reconstruction of the correct species tree with high probability. We
also validate those theoretical bounds in a simulation study. Our results
indicate that ASTRAL requires $\mathcal{O}(f^{-2} \log n)$ gene trees to
reconstruct the species tree correctly with high probability where n is the
number of species and f is the length of the shortest branch in the species
tree. Our simulations, which are the first to test ASTRAL explicitly under the
anomaly zone, show trends consistent with the theoretical bounds and also
provide some practical insights on the conditions where ASTRAL works well.
| 1 | 0 | 1 | 1 | 0 | 0 |
Two weight Commutators in the Dirichlet and Neumann Laplacian settings | In this paper we establish the characterization of the weighted BMO via two
weight commutators in the settings of the Neumann Laplacian $\Delta_{N_+}$ on
the upper half space $\mathbb{R}^n_+$ and the reflection Neumann Laplacian
$\Delta_N$ on $\mathbb{R}^n$ with respect to the weights associated to
$\Delta_{N_+}$ and $\Delta_{N}$ respectively. This in turn yields a weak
factorization for the corresponding weighted Hardy spaces, where in particular,
the weighted class associated to $\Delta_{N}$ is strictly larger than the
Muckenhoupt weighted class and contains non-doubling weights. In our study, we
also make contributions to the classical Muckenhoupt--Wheeden weighted Hardy
space (BMO space respectively) by showing that it can be characterized via area
function (Carleson measure respectively) involving the semigroup generated by
the Laplacian on $\mathbb{R}^n$ and that the duality of these weighted Hardy
and BMO spaces holds for Muckenhoupt $A^p$ weights with $p\in (1,2]$ while the
previously known related results cover only $p\in (1,{n+1\over n}]$. We also
point out that this two weight commutator theorem might not be true in the
setting of general operators $L$, and in particular we show that it is not true
when $L$ is the Dirichlet Laplacian $\Delta_{D_+}$ on $\mathbb{R}^n_+$.
| 0 | 0 | 1 | 0 | 0 | 0 |
Schoenberg Representations and Gramian Matrices of Matérn Functions | We represent Matérn functions in terms of Schoenberg's integrals which
ensure the positive definiteness and prove the systems of translates of
Matérn functions form Riesz sequences in $L^2(\R^n)$ or Sobolev spaces. Our
approach is based on a new class of integral transforms that generalize Fourier
transforms for radial functions. We also consider inverse multi-quadrics and
obtain similar results.
| 0 | 0 | 1 | 0 | 0 | 0 |
A generalization of the Hasse-Witt matrix of a hypersurface | The Hasse-Witt matrix of a hypersurface in ${\mathbb P}^n$ over a finite
field of characteristic $p$ gives essentially complete mod $p$ information
about the zeta function of the hypersurface. But if the degree $d$ of the
hypersurface is $\leq n$, the zeta function is trivial mod $p$ and the
Hasse-Witt matrix is zero-by-zero. We generalize a classical formula for the
Hasse-Witt matrix to obtain a matrix that gives a nontrivial congruence for the
zeta function for all $d$. We also describe the differential equations
satisfied by this matrix and prove that it is generically invertible.
| 0 | 0 | 1 | 0 | 0 | 0 |
Few-Shot Learning with Graph Neural Networks | We propose to study the problem of few-shot learning with the prism of
inference on a partially observed graphical model, constructed from a
collection of input images whose label can be either observed or not. By
assimilating generic message-passing inference algorithms with their
neural-network counterparts, we define a graph neural network architecture that
generalizes several of the recently proposed few-shot learning models. Besides
providing improved numerical performance, our framework is easily extended to
variants of few-shot learning, such as semi-supervised or active learning,
demonstrating the ability of graph-based models to operate well on 'relational'
tasks.
| 1 | 0 | 0 | 1 | 0 | 0 |
High-precision measurement of the proton's atomic mass | We report on the precise measurement of the atomic mass of a single proton
with a purpose-built Penning-trap system. With a precision of 32
parts-per-trillion our result not only improves on the current CODATA
literature value by a factor of three, but also disagrees with it at a level of
about 3 standard deviations.
| 0 | 1 | 0 | 0 | 0 | 0 |
Prospects of detecting HI using redshifted 21 cm radiation at z ~ 3 | Distribution of cold gas in the post-reionization era provides an important
link between distribution of galaxies and the process of star formation.
Redshifted 21 cm radiation from the Hyperfine transition of neutral Hydrogen
allows us to probe the neutral component of cold gas, most of which is to be
found in the interstellar medium of galaxies. Existing and upcoming radio
telescopes can probe the large scale distribution of neutral Hydrogen via HI
intensity mapping. In this paper we use an estimate of the HI power spectrum
derived using an ansatz to compute the expected signal from the large scale HI
distribution at z ~ 3. We find that the scale dependence of bias at small
scales makes a significant difference to the expected signal even at large
angular scales. We compare the predicted signal strength with the sensitivity
of radio telescopes that can observe such radiation and calculate the
observation time required for detecting neutral Hydrogen at these redshifts. We
find that OWFA (Ooty Wide Field Array) offers the best possibility to detect
neutral Hydrogen at z ~ 3 before the SKA (Square Kilometer Array) becomes
operational. We find that the OWFA should be able to make a 3 sigma or a more
significant detection in 2000 hours of observations at several angular scales.
Calculations done using the Fisher matrix approach indicate that a 5 sigma
detection of the binned HI power spectrum via measurement of the amplitude of
the HI power spectrum is possible in 1000 hours (Sarkar, Bharadwaj and Ali,
2017).
| 0 | 1 | 0 | 0 | 0 | 0 |
Unconditional bases of subspaces related to non-self-adjoint perturbations of self-adjoint operators | Assume that $T$ is a self-adjoint operator on a Hilbert space $\mathcal{H}$
and that the spectrum of $T$ is confined in the union $\bigcup_{j\in
J}\Delta_j$, $J\subseteq\mathbb{Z}$, of segments $\Delta_j=[\alpha_j,
\beta_j]\subset\mathbb{R}$ such that $\alpha_{j+1}>\beta_j$ and $$ \inf_{j}
\left(\alpha_{j+1}-\beta_j\right) = d > 0. $$ If $B$ is a bounded (in general
non-self-adjoint) perturbation of $T$ with $\|B\|=:b<d/2$ then the spectrum of
the perturbed operator $A=T+B$ lies in the union $\bigcup_{j\in J}
U_{b}(\Delta_j)$ of the mutually disjoint closed $b$-neighborhoods
$U_{b}(\Delta_j)$ of the segments $\Delta_j$ in $\mathbb{C}$. Let $Q_j$ be the
Riesz projection onto the invariant subspace of $A$ corresponding to the part
of the spectrum of $A$ lying in $U_{b}\left(\Delta_j\right)$, $j\in J$. Our
main result is as follows: The subspaces $\mathcal{L}_j=Q_j(\mathcal H)$, $j\in
J$, form an unconditional basis in the whole space $\mathcal H$.
| 0 | 0 | 1 | 0 | 0 | 0 |
Centralities of Nodes and Influences of Layers in Large Multiplex Networks | We formulate and propose an algorithm (MultiRank) for the ranking of nodes
and layers in large multiplex networks. MultiRank takes into account the full
multiplex network structure of the data and exploits the dual nature of the
network in terms of nodes and layers. The proposed centrality of the layers
(influences) and the centrality of the nodes are determined by a coupled set of
equations. The basic idea consists in assigning more centrality to nodes that
receive links from highly influential layers and from already central nodes.
The layers are more influential if highly central nodes are active in them. The
algorithm applies to directed/undirected as well as to weighted/unweighted
multiplex networks. We discuss the application of MultiRank to three major
examples of multiplex network datasets: the European Air Transportation
Multiplex Network, the Pierre Auger Multiplex Collaboration Network and the FAO
Multiplex Trade Network.
| 1 | 1 | 0 | 0 | 0 | 0 |
Twofold triple systems with cyclic 2-intersecting Gray codes | Given a combinatorial design $\mathcal{D}$ with block set $\mathcal{B}$, the
block-intersection graph (BIG) of $\mathcal{D}$ is the graph that has
$\mathcal{B}$ as its vertex set, where two vertices $B_{1} \in \mathcal{B}$ and
$B_{2} \in \mathcal{B} $ are adjacent if and only if $|B_{1} \cap B_{2}| > 0$.
The $i$-block-intersection graph ($i$-BIG) of $\mathcal{D}$ is the graph that
has $\mathcal{B}$ as its vertex set, where two vertices $B_{1} \in \mathcal{B}$
and $B_{2} \in \mathcal{B}$ are adjacent if and only if $|B_{1} \cap B_{2}| =
i$. In this paper several constructions are obtained that start with twofold
triple systems (TTSs) with Hamiltonian $2$-BIGs and result in larger TTSs that
also have Hamiltonian $2$-BIGs. These constructions collectively enable us to
determine the complete spectrum of TTSs with Hamiltonian $2$-BIGs (equivalently
TTSs with cyclic $2$-intersecting Gray codes) as well as the complete spectrum
for TTSs with $2$-BIGs that have Hamilton paths (i.e., for TTSs with
$2$-intersecting Gray codes).
In order to prove these spectrum results, we sometimes require ingredient
TTSs that have large partial parallel classes; we prove lower bounds on the
sizes of partial parallel clasess in arbitrary TTSs, and then construct larger
TTSs with both cyclic $2$-intersecting Gray codes and parallel classes.
| 0 | 0 | 1 | 0 | 0 | 0 |
Consequences of Unhappiness While Developing Software | The growing literature on affect among software developers mostly reports on
the linkage between happiness, software quality, and developer productivity.
Understanding the positive side of happiness -- positive emotions and moods --
is an attractive and important endeavor. Scholars in industrial and
organizational psychology have suggested that also studying the negative side
-- unhappiness -- could lead to cost-effective ways of enhancing working
conditions, job performance, and to limiting the occurrence of psychological
disorders. Our comprehension of the consequences of (un)happiness among
developers is still too shallow, and is mainly expressed in terms of
development productivity and software quality. In this paper, we attempt to
uncover the experienced consequences of unhappiness among software developers.
Using qualitative data analysis of the responses given by 181 questionnaire
participants, we identified 49 consequences of unhappiness while doing software
development. We found detrimental consequences on developers' mental
well-being, the software development process, and the produced artifacts. Our
classification scheme, available as open data, will spawn new happiness
research opportunities of cause-effect type, and it can act as a guideline for
practitioners for identifying damaging effects of unhappiness and for fostering
happiness on the job.
| 1 | 0 | 0 | 0 | 0 | 0 |
Typesafe Abstractions for Tensor Operations | We propose a typesafe abstraction to tensors (i.e. multidimensional arrays)
exploiting the type-level programming capabilities of Scala through
heterogeneous lists (HList), and showcase typesafe abstractions of common
tensor operations and various neural layers such as convolution or recurrent
neural networks. This abstraction could lay the foundation of future typesafe
deep learning frameworks that runs on Scala/JVM.
| 1 | 0 | 0 | 0 | 0 | 0 |
Structurally Sparsified Backward Propagation for Faster Long Short-Term Memory Training | Exploiting sparsity enables hardware systems to run neural networks faster
and more energy-efficiently. However, most prior sparsity-centric optimization
techniques only accelerate the forward pass of neural networks and usually
require an even longer training process with iterative pruning and retraining.
We observe that artificially inducing sparsity in the gradients of the gates in
an LSTM cell has little impact on the training quality. Further, we can enforce
structured sparsity in the gate gradients to make the LSTM backward pass up to
45% faster than the state-of-the-art dense approach and 168% faster than the
state-of-the-art sparsifying method on modern GPUs. Though the structured
sparsifying method can impact the accuracy of a model, this performance gap can
be eliminated by mixing our sparse training method and the standard dense
training method. Experimental results show that the mixed method can achieve
comparable results in a shorter time span than using purely dense training.
| 0 | 0 | 0 | 1 | 0 | 0 |
Linear density-based clustering with a discrete density model | Density-based clustering techniques are used in a wide range of data mining
applications. One of their most attractive features con- sists in not making
use of prior knowledge of the number of clusters that a dataset contains along
with their shape. In this paper we propose a new algorithm named Linear DBSCAN
(Lin-DBSCAN), a simple approach to clustering inspired by the density model
introduced with the well known algorithm DBSCAN. Designed to minimize the
computational cost of density based clustering on geospatial data, Lin-DBSCAN
features a linear time complexity that makes it suitable for real-time
applications on low-resource devices. Lin-DBSCAN uses a discrete version of the
density model of DBSCAN that takes ad- vantage of a grid-based scan and merge
approach. The name of the algorithm stems exactly from its main features
outlined above. The algorithm was tested with well known data sets.
Experimental results prove the efficiency and the validity of this approach
over DBSCAN in the context of spatial data clustering, enabling the use of a
density-based clustering technique on large datasets with low computational
cost.
| 0 | 0 | 0 | 1 | 0 | 0 |
An inexact subsampled proximal Newton-type method for large-scale machine learning | We propose a fast proximal Newton-type algorithm for minimizing regularized
finite sums that returns an $\epsilon$-suboptimal point in
$\tilde{\mathcal{O}}(d(n + \sqrt{\kappa d})\log(\frac{1}{\epsilon}))$ FLOPS,
where $n$ is number of samples, $d$ is feature dimension, and $\kappa$ is the
condition number. As long as $n > d$, the proposed method is more efficient
than state-of-the-art accelerated stochastic first-order methods for non-smooth
regularizers which requires $\tilde{\mathcal{O}}(d(n + \sqrt{\kappa
n})\log(\frac{1}{\epsilon}))$ FLOPS. The key idea is to form the subsampled
Newton subproblem in a way that preserves the finite sum structure of the
objective, thereby allowing us to leverage recent developments in stochastic
first-order methods to solve the subproblem. Experimental results verify that
the proposed algorithm outperforms previous algorithms for $\ell_1$-regularized
logistic regression on real datasets.
| 1 | 0 | 0 | 1 | 0 | 0 |
Future Energy Consumption Prediction Based on Grey Forecast Model | We use grey forecast model to predict the future energy consumption of four
states in the U.S, and make some improvments to the model.
| 0 | 0 | 0 | 1 | 0 | 0 |
AutoPass: An Automatic Password Generator | Text password has long been the dominant user authentication technique and is
used by large numbers of Internet services. If they follow recommended
practice, users are faced with the almost insuperable problem of generating and
managing a large number of site-unique and strong (i.e. non-guessable)
passwords. One way of addressing this problem is through the use of a password
generator, i.e. a client-side scheme which generates (and regenerates)
site-specific strong passwords on demand, with the minimum of user input. This
paper provides a detailed specification and analysis of AutoPass, a password
generator scheme previously outlined as part of a general analysis of such
schemes. AutoPass has been designed to address issues identified in previously
proposed password generators, and incorporates novel techniques to address
these issues. Unlike almost all previously proposed schemes, AutoPass enables
the generation of passwords that meet important real-world requirements,
including forced password changes, use of pre-specified passwords, and
generation of passwords meeting site-specific requirements.
| 1 | 0 | 0 | 0 | 0 | 0 |
A Practically Competitive and Provably Consistent Algorithm for Uplift Modeling | Randomized experiments have been critical tools of decision making for
decades. However, subjects can show significant heterogeneity in response to
treatments in many important applications. Therefore it is not enough to simply
know which treatment is optimal for the entire population. What we need is a
model that correctly customize treatment assignment base on subject
characteristics. The problem of constructing such models from randomized
experiments data is known as Uplift Modeling in the literature. Many algorithms
have been proposed for uplift modeling and some have generated promising
results on various data sets. Yet little is known about the theoretical
properties of these algorithms. In this paper, we propose a new tree-based
ensemble algorithm for uplift modeling. Experiments show that our algorithm can
achieve competitive results on both synthetic and industry-provided data. In
addition, by properly tuning the "node size" parameter, our algorithm is proved
to be consistent under mild regularity conditions. This is the first consistent
algorithm for uplift modeling that we are aware of.
| 1 | 0 | 0 | 1 | 0 | 0 |
The cosmic shoreline: the evidence that escape determines which planets have atmospheres, and what this may mean for Proxima Centauri b | The planets of the Solar System divide neatly between those with atmospheres
and those without when arranged by insolation ($I$) and escape velocity
($v_{\mathrm{esc}}$). The dividing line goes as $I \propto v_{\mathrm{esc}}^4$.
Exoplanets with reported masses and radii are shown to crowd against the
extrapolation of the Solar System trend, making a metaphorical cosmic shoreline
that unites all the planets. The $I \propto v_{\mathrm{esc}}^4$ relation may
implicate thermal escape. We therefore address the general behavior of
hydrodynamic thermal escape models ranging from Pluto to highly-irradiated
Extrasolar Giant Planets (EGPs). Energy-limited escape is harder to test
because copious XUV radiation is mostly a feature of young stars, and hence
requires extrapolating to historic XUV fluences ($I_{\mathrm{xuv}}$) using
proxies and power laws. An energy-limited shoreline should scale as
$I_{\mathrm{xuv}} \propto v_{\mathrm{esc}}^3\sqrt{\rho}$, which differs
distinctly from the apparent $I_{\mathrm{xuv}} \propto v_{\mathrm{esc}}^4$
relation. Energy-limited escape does provide good quantitative agreement to the
highly irradiated EGPs. Diffusion-limited escape implies that no planet can
lose more than 1% of its mass as H$_2$. Impact erosion, to the extent that
impact velocities $v_{\mathrm{imp}}$ can be estimated for exoplanets, fits to a
$v_{\mathrm{imp}} \approx 4\,-\,5\, v_{\mathrm{esc}}$ shoreline. The
proportionality constant is consistent with what the collision of comet
Shoemaker-Levy 9 showed us we should expect of modest impacts in deep
atmospheres. With respect to the shoreline, Proxima Centauri b is on the
metaphorical beach. Known hazards include its rapid energetic accretion, high
impact velocities, its early life on the wrong side of the runaway greenhouse,
and Proxima Centauri's XUV radiation. In its favor is a vast phase space of
unknown unknowns.
| 0 | 1 | 0 | 0 | 0 | 0 |
Universal kinetics for engagement of mechanosensing pathways in cell adhesion | When plated onto substrates, cell morphology and even stem cell
differentiation are influenced by the stiffness of their environment. Stiffer
substrates give strongly spread (eventually polarized) cells with strong focal
adhesions, and stress fibers; very soft substrates give a less developed
cytoskeleton, and much lower cell spreading. The kinetics of this process of
cell spreading is studied extensively, and important universal relationships
are established on how the cell area grows with time. Here we study the
population dynamics of spreading cells, investigating the characteristic
processes involved in cell response to the substrate. We show that unlike the
individual cell morphology, this population dynamics does not depend on the
substrate stiffness. Instead, a strong activation temperature dependence is
observed. Different cell lines on different substrates all have long-time
statistics controlled by the thermal activation over a single energy barrier
dG=19 kcal/mol, while the early-time kinetics follows a power law $t^5$. This
implies that the rate of spreading depends on an internal process of
adhesion-mechanosensing complex assembly and activation: the operational
complex must have 5 component proteins, and the last process in the sequence
(which we believe is the activation of focal adhesion kinase) is controlled by
the binding energy dG.
| 0 | 0 | 0 | 0 | 1 | 0 |
Agent-based computing from multi-agent systems to agent-based Models: a visual survey | Agent-Based Computing is a diverse research domain concerned with the
building of intelligent software based on the concept of "agents". In this
paper, we use Scientometric analysis to analyze all sub-domains of agent-based
computing. Our data consists of 1,064 journal articles indexed in the ISI web
of knowledge published during a twenty year period: 1990-2010. These were
retrieved using a topic search with various keywords commonly used in
sub-domains of agent-based computing. In our proposed approach, we have
employed a combination of two applications for analysis, namely Network
Workbench and CiteSpace - wherein Network Workbench allowed for the analysis of
complex network aspects of the domain, detailed visualization-based analysis of
the bibliographic data was performed using CiteSpace. Our results include the
identification of the largest cluster based on keywords, the timeline of
publication of index terms, the core journals and key subject categories. We
also identify the core authors, top countries of origin of the manuscripts
along with core research institutes. Finally, our results have interestingly
revealed the strong presence of agent-based computing in a number of
non-computing related scientific domains including Life Sciences, Ecological
Sciences and Social Sciences.
| 1 | 1 | 0 | 0 | 0 | 0 |
Large Spontaneous Hall Effects in Chiral Topological Magnets | As novel topological phases in correlated electron systems, we have found two
examples of non-ferromagnetic states that exhibit a large anomalous Hall
effect. One is the chiral spin liquid compound Pr$_{2}$Ir$_{2}$O$_{7}$, which
exhibits a spontaneous Hall effect in a spin liquid state due to spin ice
correlation. The other is the chiral antiferromagnets Mn$_{3}$Sn and Mn$_{3}$Ge
that exhibit a large anomalous Hall effect at room temperature. The latter
shows a sign change of the anomalous Hall effect by a small change in the
magnetic field by a few 100 G, which should be useful for various applications.
We will discuss that the magnetic Weyl metal states are the origin for such a
large anomalous Hall effect observed in both the spin liquid and
antiferromagnet that possess almost no magnetization.
| 0 | 1 | 0 | 0 | 0 | 0 |
Mott metal-insulator transition in the Doped Hubbard-Holstein model | Motivated by the current interest in the understanding of the Mott insulators
away from half filling, observed in many perovskite oxides, we study the Mott
metal-insulator transition (MIT) in the doped Hubbard-Holstein model using the
Hatree-Fock mean field theory. The Hubbard-Holstein model is the simplest model
containing both the Coulomb and the electron-lattice interactions, which are
important ingredients in the physics of the perovskite oxides. In contrast to
the half-filled Hubbard model, which always results in a single phase (either
metallic or insulating), our results show that away from half-filling, a mixed
phase of metallic and insulating regions occur. As the dopant concentration is
increased, the metallic part progressively grows in volume, until it exceeds
the percolation threshold, leading to percolative conduction. This happens
above a critical dopant concentration $\delta_c$, which, depending on the
strength of the electron-lattice interaction, can be a significant fraction of
unity. This means that the material could be insulating even for a substantial
amount of doping, in contrast to the expectation that doped holes would destroy
the insulating behavior of the half-filled Hubbard model. Our theory provides a
framework for the understanding of the density-driven metal-insulator
transition observed in many complex oxides.
| 0 | 1 | 0 | 0 | 0 | 0 |
ASDA : Analyseur Syntaxique du Dialecte Alg{é}rien dans un but d'analyse s{é}mantique | Opinion mining and sentiment analysis in social media is a research issue
having a great interest in the scientific community. However, before begin this
analysis, we are faced with a set of problems. In particular, the problem of
the richness of languages and dialects within these media. To address this
problem, we propose in this paper an approach of construction and
implementation of Syntactic analyzer named ASDA. This tool represents a parser
for the Algerian dialect that label the terms of a given corpus. Thus, we
construct a labeling table containing for each term its stem, different
prefixes and suffixes, allowing us to determine the different grammatical parts
a sort of POS tagging. This labeling will serve us later in the semantic
processing of the Algerian dialect, like the automatic translation of this
dialect or sentiment analysis
| 1 | 0 | 0 | 0 | 0 | 0 |
Latent Intention Dialogue Models | Developing a dialogue agent that is capable of making autonomous decisions
and communicating by natural language is one of the long-term goals of machine
learning research. Traditional approaches either rely on hand-crafting a small
state-action set for applying reinforcement learning that is not scalable or
constructing deterministic models for learning dialogue sentences that fail to
capture natural conversational variability. In this paper, we propose a Latent
Intention Dialogue Model (LIDM) that employs a discrete latent variable to
learn underlying dialogue intentions in the framework of neural variational
inference. In a goal-oriented dialogue scenario, these latent intentions can be
interpreted as actions guiding the generation of machine responses, which can
be further refined autonomously by reinforcement learning. The experimental
evaluation of LIDM shows that the model out-performs published benchmarks for
both corpus-based and human evaluation, demonstrating the effectiveness of
discrete latent variable models for learning goal-oriented dialogues.
| 1 | 0 | 0 | 1 | 0 | 0 |
Quasiconvex elastodynamics: weak-strong uniqueness for measure-valued solutions | A weak-strong uniqueness result is proved for measure-valued solutions to the
system of conservation laws arising in elastodynamics. The main novelty brought
forward by the present work is that the underlying stored-energy function of
the material is assumed strongly quasiconvex. The proof employs tools from the
calculus of variations to establish general convexity-type bounds on
quasiconvex functions and recasts them in order to adapt the relative entropy
method to quasiconvex elastodynamics.
| 0 | 0 | 1 | 0 | 0 | 0 |
Accelerated Dual Learning by Homotopic Initialization | Gradient descent and coordinate descent are well understood in terms of their
asymptotic behavior, but less so in a transient regime often used for
approximations in machine learning. We investigate how proper initialization
can have a profound effect on finding near-optimal solutions quickly. We show
that a certain property of a data set, namely the boundedness of the
correlations between eigenfeatures and the response variable, can lead to
faster initial progress than expected by commonplace analysis. Convex
optimization problems can tacitly benefit from that, but this automatism does
not apply to their dual formulation. We analyze this phenomenon and devise
provably good initialization strategies for dual optimization as well as
heuristics for the non-convex case, relevant for deep learning. We find our
predictions and methods to be experimentally well-supported.
| 1 | 0 | 0 | 0 | 0 | 0 |
Inverse Reinforcement Learning from Summary Data | Inverse reinforcement learning (IRL) aims to explain observed strategic
behavior by fitting reinforcement learning models to behavioral data. However,
traditional IRL methods are only applicable when the observations are in the
form of state-action paths. This assumption may not hold in many real-world
modeling settings, where only partial or summarized observations are available.
In general, we may assume that there is a summarizing function $\sigma$, which
acts as a filter between us and the true state-action paths that constitute the
demonstration. Some initial approaches to extending IRL to such situations have
been presented, but with very specific assumptions about the structure of
$\sigma$, such as that only certain state observations are missing. This paper
instead focuses on the most general case of the problem, where no assumptions
are made about the summarizing function, except that it can be evaluated. We
demonstrate that inference is still possible. The paper presents exact and
approximate inference algorithms that allow full posterior inference, which is
particularly important for assessing parameter uncertainty in this challenging
inference situation. Empirical scalability is demonstrated to reasonably sized
problems, and practical applicability is demonstrated by estimating the
posterior for a cognitive science RL model based on an observed user's task
completion time only.
| 1 | 0 | 0 | 1 | 0 | 0 |
Algorithm for Optimization and Interpolation based on Hyponormality | On one hand, consider the problem of finding global solutions to a polynomial
optimization problem and, on the other hand, consider the problem of
interpolating a set of points with a complex exponential function. This paper
proposes a single algorithm to address both problems. It draws on the notion of
hyponormality in operator theory. Concerning optimization, it seems to be the
first algorithm that is capable of extracting global solutions from a
polynomial optimization problem where the variables and data are complex
numbers. It also applies to real polynomial optimization, a special case of
complex polynomial optimization, and thus extends the work of Henrion and
Lasserre implemented in GloptiPoly. Concerning interpolation, the algorithm
provides an alternative to Prony's method based on the Autonne-Takagi
factorization and it avoids solving a Vandermonde system. The algorithm and its
proof are based exclusively on linear algebra. They are devoid of notions from
algebraic geometry, contrary to existing methods for interpolation. The
algorithm is tested on a series of examples, each illustrating a different
facet of the approach. One of the examples demonstrates that hyponormality can
be enforced numerically to strenghten a convex relaxation and to force its
solution to have rank one.
| 0 | 0 | 1 | 0 | 0 | 0 |
Human experts vs. machines in taxa recognition | The step of expert taxa recognition currently slows down the response time of
many bioassessments. Shifting to quicker and cheaper state-of-the-art machine
learning approaches is still met with expert scepticism towards the ability and
logic of machines. In our study, we investigate both the differences in
accuracy and in the identification logic of taxonomic experts and machines. We
propose a systematic approach utilizing deep Convolutional Neural Nets with the
transfer learning paradigm and extensively evaluate it over a multi-label and
multi-pose taxonomic dataset specifically created for this comparison. We also
study the prediction accuracy on different ranks of taxonomic hierarchy in
detail. Our results revealed that human experts using actual specimens yield
the lowest classification error. However, our proposed, much faster, automated
approach using deep Convolutional Neural Nets comes very close to human
accuracy. Contrary to previous findings in the literature, we find that
machines following a typical flat classification approach commonly used in
machine learning performs better than forcing machines to adopt a hierarchical,
local per parent node approach used by human taxonomic experts. Finally, we
publicly share our unique dataset to serve as a public benchmark dataset in
this field.
| 1 | 0 | 0 | 1 | 0 | 0 |
A micrometer-thick oxide film with high thermoelectric performance at temperature ranging from 20-400 K | Thermoelectric (TE) materials achieve localised conversion between thermal
and electric energies, and the conversion efficiency is determined by a figure
of merit zT. Up to date, two-dimensional electron gas (2DEG) related TE
materials hold the records for zT near room-temperature. A sharp increase in zT
up to ~2.0 was observed previously for superlattice materials such as PbSeTe,
Bi2Te3/Sb2Te3 and SrNb0.2Ti0.8O3/SrTiO3, when the thicknesses of these TE
materials were spatially confine within sub-nanometre scale. The
two-dimensional confinement of carriers enlarges the density of states near the
Fermi energy3-6 and triggers electron phonon coupling. This overcomes the
conventional {\sigma}-S trade-off to more independently improve S, and thereby
further increases thermoelectric power factors (PF=S2{\sigma}). Nevertheless,
practical applications of the present 2DEG materials for high power energy
conversions are impeded by the prerequisite of spatial confinement, as the
amount of TE material is insufficient. Here, we report similar TE properties to
2DEGs but achieved in SrNb0.2Ti0.8O3 films with thickness within sub-micrometer
scale by regulating interfacial and lattice polarizations. High power factor
(up to 103 {\mu}Wcm-1K-2) and zT value (up to 1.6) were observed for the film
materials near room-temperature and below. Even reckon in the thickness of the
substrate, an integrated power factor of both film and substrate approaching to
be 102 {\mu}Wcm-1K-2 was achieved in a 2 {\mu}m-thick SrNb0.2Ti0.8O3 film grown
on a 100 {\mu}m-thick SrTiO3 substrate. The dependence of high TE performances
on size-confinement is reduced by ~103 compared to the conventional
2DEG-related TE materials. As-grown oxide films are less toxic and not
dependent on large amounts of heavy elements, potentially paving the way
towards applications in localised refrigeration and electric power generations.
| 0 | 1 | 0 | 0 | 0 | 0 |
MIT at SemEval-2017 Task 10: Relation Extraction with Convolutional Neural Networks | Over 50 million scholarly articles have been published: they constitute a
unique repository of knowledge. In particular, one may infer from them
relations between scientific concepts, such as synonyms and hyponyms.
Artificial neural networks have been recently explored for relation extraction.
In this work, we continue this line of work and present a system based on a
convolutional neural network to extract relations. Our model ranked first in
the SemEval-2017 task 10 (ScienceIE) for relation extraction in scientific
articles (subtask C).
| 1 | 0 | 0 | 1 | 0 | 0 |
Maximizing acquisition functions for Bayesian optimization | Bayesian optimization is a sample-efficient approach to global optimization
that relies on theoretically motivated value heuristics (acquisition functions)
to guide its search process. Fully maximizing acquisition functions produces
the Bayes' decision rule, but this ideal is difficult to achieve since these
functions are frequently non-trivial to optimize. This statement is especially
true when evaluating queries in parallel, where acquisition functions are
routinely non-convex, high-dimensional, and intractable. We first show that
acquisition functions estimated via Monte Carlo integration are consistently
amenable to gradient-based optimization. Subsequently, we identify a common
family of acquisition functions, including EI and UCB, whose properties not
only facilitate but justify use of greedy approaches for their maximization.
| 0 | 0 | 0 | 1 | 0 | 0 |
Angular momentum evolution of galaxies over the past 10-Gyr: A MUSE and KMOS dynamical survey of 400 star-forming galaxies from z=0.3-1.7 | We present a MUSE and KMOS dynamical study 405 star-forming galaxies at
redshift z=0.28-1.65 (median redshift z=0.84). Our sample are representative of
star-forming, main-sequence galaxies, with star-formation rates of
SFR=0.1-30Mo/yr and stellar masses M=10^8-10^11Mo. For 49+/-4% of our sample,
the dynamics suggest rotational support, 24+/-3% are unresolved systems and
5+/-2% appear to be early-stage major mergers with components on 8-30kpc
scales. The remaining 22+/-5% appear to be dynamically complex, irregular (or
face-on systems). For galaxies whose dynamics suggest rotational support, we
derive inclination corrected rotational velocities and show these systems lie
on a similar scaling between stellar mass and specific angular momentum as
local spirals with j*=J/M*\propto M^(2/3) but with a redshift evolution that
scales as j*\propto M^{2/3}(1+z)^(-1). We identify a correlation between
specific angular momentum and disk stability such that galaxies with the
highest specific angular momentum, log(j*/M^(2/3))>2.5, are the most stable,
with Toomre Q=1.10+/-0.18, compared to Q=0.53+/-0.22 for galaxies with
log(j*/M^(2/3))<2.5. At a fixed mass, the HST morphologies of galaxies with the
highest specific angular momentum resemble spiral galaxies, whilst those with
low specific angular momentum are morphologically complex and dominated by
several bright star-forming regions. This suggests that angular momentum plays
a major role in defining the stability of gas disks: at z~1, massive galaxies
that have disks with low specific angular momentum, appear to be globally
unstable, clumpy and turbulent systems. In contrast, galaxies with high
specific angular have evolved in to stable disks with spiral structures.
| 0 | 1 | 0 | 0 | 0 | 0 |
Iterative Object and Part Transfer for Fine-Grained Recognition | The aim of fine-grained recognition is to identify sub-ordinate categories in
images like different species of birds. Existing works have confirmed that, in
order to capture the subtle differences across the categories, automatic
localization of objects and parts is critical. Most approaches for object and
part localization relied on the bottom-up pipeline, where thousands of region
proposals are generated and then filtered by pre-trained object/part models.
This is computationally expensive and not scalable once the number of
objects/parts becomes large. In this paper, we propose a nonparametric
data-driven method for object and part localization. Given an unlabeled test
image, our approach transfers annotations from a few similar images retrieved
in the training set. In particular, we propose an iterative transfer strategy
that gradually refine the predicted bounding boxes. Based on the located
objects and parts, deep convolutional features are extracted for recognition.
We evaluate our approach on the widely-used CUB200-2011 dataset and a new and
large dataset called Birdsnap. On both datasets, we achieve better results than
many state-of-the-art approaches, including a few using oracle (manually
annotated) bounding boxes in the test images.
| 1 | 0 | 0 | 0 | 0 | 0 |
On measures of edge-uncolorability of cubic graphs: A brief survey and some new results | There are many hard conjectures in graph theory, like Tutte's 5-flow
conjecture, and the 5-cycle double cover conjecture, which would be true in
general if they would be true for cubic graphs. Since most of them are
trivially true for 3-edge-colorable cubic graphs, cubic graphs which are not
3-edge-colorable, often called {\em snarks}, play a key role in this context.
Here, we survey parameters measuring how far apart a non 3-edge-colorable graph
is from being 3-edge-colorable. We study their interrelation and prove some new
results. Besides getting new insight into the structure of snarks, we show that
such measures give partial results with respect to these important conjectures.
The paper closes with a list of open problems and conjectures.
| 0 | 0 | 1 | 0 | 0 | 0 |
Gas around galaxy haloes - III: hydrogen absorption signatures around galaxies and QSOs in the Sherwood simulation suite | Modern theories of galaxy formation predict that galaxies impact on their
gaseous surroundings, playing the fundamental role of regulating the amount of
gas converted into stars. While star-forming galaxies are believed to provide
feedback through galactic winds, Quasi-Stellar Objects (QSOs) are believed
instead to provide feedback through the heat generated by accretion onto a
central supermassive black hole. A quantitative difference in the impact of
feedback on the gaseous environments of star-forming galaxies and QSOs has not
been established through direct observations. Using the Sherwood cosmological
simulations, we demonstrate that measurements of neutral hydrogen in the
vicinity of star-forming galaxies and QSOs during the era of peak galaxy
formation show excess LyA absorption extending up to comoving radii of about
150 kpc for star-forming galaxies and 300 - 700 kpc for QSOs. Simulations
including supernovae-driven winds with the wind velocity scaling like the
escape velocity of the halo account for the absorption around star-forming
galaxies but not QSOs.
| 0 | 1 | 0 | 0 | 0 | 0 |
Distributed Newton Methods for Deep Neural Networks | Deep learning involves a difficult non-convex optimization problem with a
large number of weights between any two adjacent layers of a deep structure. To
handle large data sets or complicated networks, distributed training is needed,
but the calculation of function, gradient, and Hessian is expensive. In
particular, the communication and the synchronization cost may become a
bottleneck. In this paper, we focus on situations where the model is
distributedly stored, and propose a novel distributed Newton method for
training deep neural networks. By variable and feature-wise data partitions,
and some careful designs, we are able to explicitly use the Jacobian matrix for
matrix-vector products in the Newton method. Some techniques are incorporated
to reduce the running time as well as the memory consumption. First, to reduce
the communication cost, we propose a diagonalization method such that an
approximate Newton direction can be obtained without communication between
machines. Second, we consider subsampled Gauss-Newton matrices for reducing the
running time as well as the communication cost. Third, to reduce the
synchronization cost, we terminate the process of finding an approximate Newton
direction even though some nodes have not finished their tasks. Details of some
implementation issues in distributed environments are thoroughly investigated.
Experiments demonstrate that the proposed method is effective for the
distributed training of deep neural networks. In compared with stochastic
gradient methods, it is more robust and may give better test accuracy.
| 0 | 0 | 0 | 1 | 0 | 0 |
Personalized advice for enhancing well-being using automated impulse response analysis --- AIRA | The attention for personalized mental health care is thriving. Research data
specific to the individual, such as time series sensor data or data from
intensive longitudinal studies, is relevant from a research perspective, as
analyses on these data can reveal the heterogeneity among the participants and
provide more precise and individualized results than with group-based methods.
However, using this data for self-management and to help the individual to
improve his or her mental health has proven to be challenging.
The present work describes a novel approach to automatically generate
personalized advice for the improvement of the well-being of individuals by
using time series data from intensive longitudinal studies: Automated Impulse
Response Analysis (AIRA). AIRA analyzes vector autoregression models of
well-being by generating impulse response functions. These impulse response
functions are used in simulations to determine which variables in the model
have the largest influence on the other variables and thus on the well-being of
the participant. The effects found can be used to support self-management.
We demonstrate the practical usefulness of AIRA by performing analysis on
longitudinal self-reported data about psychological variables. To evaluate its
effectiveness and efficacy, we ran its algorithms on two data sets ($N=4$ and
$N=5$), and discuss the results. Furthermore, we compare AIRA's output to the
results of a previously published study and show that the results are
comparable. By automating Impulse Response Function Analysis, AIRA fulfills the
need for accurate individualized models of health outcomes at a low resource
cost with the potential for upscaling.
| 1 | 0 | 0 | 0 | 0 | 0 |
Being Robust (in High Dimensions) Can Be Practical | Robust estimation is much more challenging in high dimensions than it is in
one dimension: Most techniques either lead to intractable optimization problems
or estimators that can tolerate only a tiny fraction of errors. Recent work in
theoretical computer science has shown that, in appropriate distributional
models, it is possible to robustly estimate the mean and covariance with
polynomial time algorithms that can tolerate a constant fraction of
corruptions, independent of the dimension. However, the sample and time
complexity of these algorithms is prohibitively large for high-dimensional
applications. In this work, we address both of these issues by establishing
sample complexity bounds that are optimal, up to logarithmic factors, as well
as giving various refinements that allow the algorithms to tolerate a much
larger fraction of corruptions. Finally, we show on both synthetic and real
data that our algorithms have state-of-the-art performance and suddenly make
high-dimensional robust estimation a realistic possibility.
| 1 | 0 | 0 | 1 | 0 | 0 |
Properties of In-Plane Graphene/MoS2 Heterojunctions | The graphene/MoS2 heterojunction formed by joining the two components
laterally in a single plane promises to exhibit a low-resistance contact
according to the Schottky-Mott rule. Here we provide an atomic-scale
description of the structural, electronic, and magnetic properties of this type
of junction. We first identify the energetically favorable structures in which
the preference of forming C-S or C-Mo bonds at the boundary depends on the
chemical conditions. We find that significant charge transfer between graphene
and MoS2 is localized at the boundary. We show that the abundant 1D boundary
states substantially pin the Fermi level in the lateral contact between
graphene and MoS2, in close analogy to the effect of 2D interfacial states in
the contacts between 3D materials. Furthermore, we propose specific ways in
which these effects can be exploited to achieve spin-polarized currents.
| 0 | 1 | 0 | 0 | 0 | 0 |
Semi-extraspecial groups with an abelian subgroup of maximal possible order | Let $p$ be a prime. A $p$-group $G$ is defined to be semi-extraspecial if for
every maximal subgroup $N$ in $Z(G)$ the quotient $G/N$ is a an extraspecial
group. In addition, we say that $G$ is ultraspecial if $G$ is semi-extraspecial
and $|G:G'| = |G'|^2$. In this paper, we prove that every $p$-group of
nilpotence class $2$ is isomorphic to a subgroup of some ultraspecial group.
Given a prime $p$ and a positive integer $n$, we provide a framework to
construct of all the ultraspecial groups order $p^{3n}$ that contain an abelian
subgroup of order $p^{2n}$. In the literature, it has been proved that every
ultraspecial group $G$ order $p^{3n}$ with at least two abelian subgroups of
order $p^{2n}$ can be associated to a semifield. We provide a generalization of
semifield, and then we show that every semi-extraspecial group $G$ that is the
product of two abelian subgroups can be associated with this generalization of
semifield.
| 0 | 0 | 1 | 0 | 0 | 0 |
Genetic and Memetic Algorithm with Diversity Equilibrium based on Greedy Diversification | The lack of diversity in a genetic algorithm's population may lead to a bad
performance of the genetic operators since there is not an equilibrium between
exploration and exploitation. In those cases, genetic algorithms present a fast
and unsuitable convergence.
In this paper we develop a novel hybrid genetic algorithm which attempts to
obtain a balance between exploration and exploitation. It confronts the
diversity problem using the named greedy diversification operator. Furthermore,
the proposed algorithm applies a competition between parent and children so as
to exploit the high quality visited solutions. These operators are complemented
by a simple selection mechanism designed to preserve and take advantage of the
population diversity.
Additionally, we extend our proposal to the field of memetic algorithms,
obtaining an improved model with outstanding results in practice.
The experimental study shows the validity of the approach as well as how
important is taking into account the exploration and exploitation concepts when
designing an evolutionary algorithm.
| 1 | 0 | 0 | 0 | 0 | 0 |
Determinants of Mobile Money Adoption in Pakistan | In this work, we analyze the problem of adoption of mobile money in Pakistan
by using the call detail records of a major telecom company as our input. Our
results highlight the fact that different sections of the society have
different patterns of adoption of digital financial services but user mobility
related features are the most important one when it comes to adopting and using
mobile money services.
| 0 | 0 | 0 | 1 | 0 | 0 |
Cherlin's conjecture for almost simple groups of Lie rank 1 | We prove Cherlin's conjecture, concerning binary primitive permutation
groups, for those groups with socle isomorphic to $\mathrm{PSL}_2(q)$,
${^2\mathrm{B}_2}(q)$, ${^2\mathrm{G}_2}(q)$ or $\mathrm{PSU}_3(q)$. Our method
uses the notion of a "strongly non-binary action".
| 0 | 0 | 1 | 0 | 0 | 0 |
Universal Function Approximation by Deep Neural Nets with Bounded Width and ReLU Activations | This article concerns the expressive power of depth in neural nets with ReLU
activations and bounded width. We are particularly interested in the following
questions: what is the minimal width $w_{\text{min}}(d)$ so that ReLU nets of
width $w_{\text{min}}(d)$ (and arbitrary depth) can approximate any continuous
function on the unit cube $[0,1]^d$ aribitrarily well? For ReLU nets near this
minimal width, what can one say about the depth necessary to approximate a
given function? Our approach to this paper is based on the observation that,
due to the convexity of the ReLU activation, ReLU nets are particularly
well-suited for representing convex functions. In particular, we prove that
ReLU nets with width $d+1$ can approximate any continuous convex function of
$d$ variables arbitrarily well. These results then give quantitative depth
estimates for the rate of approximation of any continuous scalar function on
the $d$-dimensional cube $[0,1]^d$ by ReLU nets with width $d+3.$
| 1 | 0 | 1 | 1 | 0 | 0 |
Network Essence: PageRank Completion and Centrality-Conforming Markov Chains | Jiří Matoušek (1963-2015) had many breakthrough contributions in
mathematics and algorithm design. His milestone results are not only profound
but also elegant. By going beyond the original objects --- such as Euclidean
spaces or linear programs --- Jirka found the essence of the challenging
mathematical/algorithmic problems as well as beautiful solutions that were
natural to him, but were surprising discoveries to the field.
In this short exploration article, I will first share with readers my initial
encounter with Jirka and discuss one of his fundamental geometric results from
the early 1990s. In the age of social and information networks, I will then
turn the discussion from geometric structures to network structures, attempting
to take a humble step towards the holy grail of network science, that is to
understand the network essence that underlies the observed
sparse-and-multifaceted network data. I will discuss a simple result which
summarizes some basic algebraic properties of personalized PageRank matrices.
Unlike the traditional transitive closure of binary relations, the personalized
PageRank matrices take "accumulated Markovian closure" of network data. Some of
these algebraic properties are known in various contexts. But I hope featuring
them together in a broader context will help to illustrate the desirable
properties of this Markovian completion of networks, and motivate systematic
developments of a network theory for understanding vast and ubiquitous
multifaceted network data.
| 1 | 0 | 0 | 1 | 0 | 0 |
A Regularized Framework for Sparse and Structured Neural Attention | Modern neural networks are often augmented with an attention mechanism, which
tells the network where to focus within the input. We propose in this paper a
new framework for sparse and structured attention, building upon a smoothed max
operator. We show that the gradient of this operator defines a mapping from
real values to probabilities, suitable as an attention mechanism. Our framework
includes softmax and a slight generalization of the recently-proposed sparsemax
as special cases. However, we also show how our framework can incorporate
modern structured penalties, resulting in more interpretable attention
mechanisms, that focus on entire segments or groups of an input. We derive
efficient algorithms to compute the forward and backward passes of our
attention mechanisms, enabling their use in a neural network trained with
backpropagation. To showcase their potential as a drop-in replacement for
existing ones, we evaluate our attention mechanisms on three large-scale tasks:
textual entailment, machine translation, and sentence summarization. Our
attention mechanisms improve interpretability without sacrificing performance;
notably, on textual entailment and summarization, we outperform the standard
attention mechanisms based on softmax and sparsemax.
| 1 | 0 | 0 | 1 | 0 | 0 |
Intrinsically Motivated Goal Exploration Processes with Automatic Curriculum Learning | Intrinsically motivated spontaneous exploration is a key enabler of
autonomous lifelong learning in human children. It allows them to discover and
acquire large repertoires of skills through self-generation, self-selection,
self-ordering and self-experimentation of learning goals. We present the
unsupervised multi-goal reinforcement learning formal framework as well as an
algorithmic approach called intrinsically motivated goal exploration processes
(IMGEP) to enable similar properties of autonomous learning in machines. The
IMGEP algorithmic architecture relies on several principles: 1) self-generation
of goals as parameterized reinforcement learning problems; 2) selection of
goals based on intrinsic rewards; 3) exploration with parameterized
time-bounded policies and fast incremental goal-parameterized policy search; 4)
systematic reuse of information acquired when targeting a goal for improving
other goals. We present a particularly efficient form of IMGEP that uses a
modular representation of goal spaces as well as intrinsic rewards based on
learning progress. We show how IMGEPs automatically generate a learning
curriculum within an experimental setup where a real humanoid robot can explore
multiple spaces of goals with several hundred continuous dimensions. While no
particular target goal is provided to the system beforehand, this curriculum
allows the discovery of skills of increasing complexity, that act as stepping
stone for learning more complex skills (like nested tool use). We show that
learning several spaces of diverse problems can be more efficient for learning
complex skills than only trying to directly learn these complex skills. We
illustrate the computational efficiency of IMGEPs as these robotic experiments
use a simple memory-based low-level policy representations and search
algorithm, enabling the whole system to learn online and incrementally on a
Raspberry Pi 3.
| 1 | 0 | 0 | 0 | 0 | 0 |
Sentiment Perception of Readers and Writers in Emoji use | Previous research has traditionally analyzed emoji sentiment from the point
of view of the reader of the content not the author. Here, we analyze emoji
sentiment from the point of view of the author and present a emoji sentiment
benchmark that was built from an employee happiness dataset where emoji happen
to be annotated with daily happiness of the author of the comment. The data
spans over 3 years, and 4k employees of 56 companies based in Barcelona. We
compare sentiment of writers to readers. Results indicate that, there is an 82%
agreement in how emoji sentiment is perceived by readers and writers. Finally,
we report that when authors use emoji they report higher levels of happiness.
Emoji use was not found to be correlated with differences in author moodiness.
| 1 | 0 | 0 | 0 | 0 | 0 |
On C-class equations | The concept of a C-class of differential equations goes back to E. Cartan
with the upshot that generic equations in a C-class can be solved without
integration. While Cartan's definition was in terms of differential invariants
being first integrals, all results exhibiting C-classes that we are aware of
are based on the fact that a canonical Cartan geometry associated to the
equations in the class descends to the space of solutions. For sufficiently low
orders, these geometries belong to the class of parabolic geometries and the
results follow from the general characterization of geometries descending to a
twistor space.
In this article we answer the question of whether a canonical Cartan geometry
descends to the space of solutions in the case of scalar ODEs of order at least
four and of systems of ODEs of order at least three. As in the lower order
cases, this is characterized by the vanishing of the generalized Wilczynski
invariants, which are defined via the linearization at a solution. The
canonical Cartan geometries (which are not parabolic geometries) are a slight
variation of those available in the literature based on a recent general
construction. All the verifications needed to apply this construction for the
classes of ODEs we study are carried out in the article, which thus also
provides a complete alternative proof for the existence of canonical Cartan
connections associated to higher order (systems of) ODEs.
| 0 | 0 | 1 | 0 | 0 | 0 |
Confidence Intervals for Quantiles from Histograms and Other Grouped Data | Interval estimation of quantiles has been treated by many in the literature.
However, to the best of our knowledge there has been no consideration for
interval estimation when the data are available in grouped format. Motivated by
this, we introduce several methods to obtain confidence intervals for quantiles
when only grouped data is available. Our preferred method for interval
estimation is to approximate the underlying density using the Generalized
Lambda Distribution (GLD) to both estimate the quantiles and variance of the
quantile estimators. We compare the GLD method with some other methods that we
also introduce which are based on a frequency approximation approach and a
linear interpolation approximation of the density. Our methods are strongly
supported by simulations showing that excellent coverage can be achieved for a
wide number of distributions. These distributions include highly-skewed
distributions such as the log-normal, Dagum and Singh-Maddala distributions. We
also apply our methods to real data and show that inference can be carried out
on published outcomes that have been summarized only by a histogram. Our
methods are therefore useful for a broad range of applications. We have also
created a web application that can be used to conveniently calculate the
estimators.
| 0 | 0 | 0 | 1 | 0 | 0 |
Reach and speed of judgment propagation in the laboratory | In recent years, a large body of research has demonstrated that judgments and
behaviors can propagate from person to person. Phenomena as diverse as
political mobilization, health practices, altruism, and emotional states
exhibit similar dynamics of social contagion. The precise mechanisms of
judgment propagation are not well understood, however, because it is difficult
to control for confounding factors such as homophily or dynamic network
structures. We introduce a novel experimental design that renders possible the
stringent study of judgment propagation. In this design, experimental chains of
individuals can revise their initial judgment in a visual perception task after
observing a predecessor's judgment. The positioning of a very good performer at
the top of a chain created a performance gap, which triggered waves of judgment
propagation down the chain. We evaluated the dynamics of judgment propagation
experimentally. Despite strong social influence within pairs of individuals,
the reach of judgment propagation across a chain rarely exceeded a social
distance of three to four degrees of separation. Furthermore, computer
simulations showed that the speed of judgment propagation decayed exponentially
with the social distance from the source. We show that information distortion
and the overweighting of other people's errors are two individual-level
mechanisms hindering judgment propagation at the scale of the chain. Our
results contribute to the understanding of social contagion processes, and our
experimental method offers numerous new opportunities to study judgment
propagation in the laboratory.
| 1 | 1 | 0 | 0 | 0 | 0 |
Texture Characterization by Using Shape Co-occurrence Patterns | Texture characterization is a key problem in image understanding and pattern
recognition. In this paper, we present a flexible shape-based texture
representation using shape co-occurrence patterns. More precisely, texture
images are first represented by tree of shapes, each of which is associated
with several geometrical and radiometric attributes. Then four typical kinds of
shape co-occurrence patterns based on the hierarchical relationship of the
shapes in the tree are learned as codewords. Three different coding methods are
investigated to learn the codewords, with which, any given texture image can be
encoded into a descriptive vector. In contrast with existing works, the
proposed method not only inherits the strong ability to depict geometrical
aspects of textures and the high robustness to variations of imaging conditions
from the shape-based method, but also provides a flexible way to consider shape
relationships and to compute high-order statistics on the tree. To our
knowledge, this is the first time to use co-occurrence patterns of explicit
shapes as a tool for texture analysis. Experiments on various texture datasets
and scene datasets demonstrate the efficiency of the proposed method.
| 1 | 0 | 0 | 0 | 0 | 0 |
NGC 3105: A Young Cluster in the Outer Galaxy | Images and spectra of the open cluster NGC 3105 have been obtained with GMOS
on Gemini South. The (i', g'-i') color-magnitude diagram (CMD) constructed from
these data extends from the brightest cluster members to g'~23. This is 4 - 5
mag fainter than previous CMDs at visible wavelengths and samples cluster
members with sub-solar masses. Assuming a half-solar metallicity, comparisons
with isochrones yield a distance of 6.6+/-0.3 kpc. An age of at least 32 Myr is
found based on the photometric properties of the brightest stars, coupled with
the apparent absence of pre-main sequence stars in the lower regions of the
CMD. The luminosity function of stars between 50 and 70 arcsec from the cluster
center is consistent with a Chabrier lognormal mass function. However, at radii
smaller than 50 arcsec there is a higher specific frequency of the most massive
main sequence stars than at larger radii. Photometry obtained from archival
SPITZER images reveals that some of the brightest stars near NGC 3105 have
excess infrared emission, presumably from warm dust envelopes. Halpha emission
is detected in a few early-type stars in and around the cluster, building upon
previous spectroscopic observations that found Be stars near NGC 3105. The
equivalent width of the NaD lines in the spectra of early type stars is
consistent with the reddening found from comparisons with isochrones. Stars
with i'~18.5 that fall near the cluster main sequence have a spectral-type A5V,
and a distance modulus that is consistent with that obtained by comparing
isochrones with the CMD is found assuming solar neighborhood intrinsic
brightnesses for these stars.
| 0 | 1 | 0 | 0 | 0 | 0 |
Exact solution of a two-species quantum dimer model for pseudogap metals | We present an exact ground state solution of a quantum dimer model introduced
in Ref.[1], which features ordinary bosonic spin-singlet dimers as well as
fermionic dimers that can be viewed as bound states of spinons and holons in a
hole-doped resonating valence bond liquid. Interestingly, this model captures
several essential properties of the metallic pseudogap phase in high-$T_c$
cuprate superconductors. We identify a line in parameter space where the exact
ground state wave functions can be constructed at an arbitrary density of
fermionic dimers. At this exactly solvable line the ground state has a huge
degeneracy, which can be interpreted as a flat band of fermionic excitations.
Perturbing around the exactly solvable line, this degeneracy is lifted and the
ground state is a fractionalized Fermi liquid with a small pocket Fermi surface
in the low doping limit.
| 0 | 1 | 0 | 0 | 0 | 0 |
Strong instability of standing waves for nonlinear Schrödinger equations with a partial confinement | We study the instability of standing wave solutions for nonlinear
Schrödinger equations with a one-dimensional harmonic potential in
dimension $N\ge 2$. We prove that if the nonlinearity is $L^2$-critical or
supercritical in dimension $N-1$, then any ground states are strongly unstable
by blowup.
| 0 | 0 | 1 | 0 | 0 | 0 |
Theory and Applications of Matrix-Weighted Consensus | This paper proposes the matrix-weighted consensus algorithm, which is a
generalization of the consensus algorithm in the literature. Given a networked
dynamical system where the interconnections between agents are weighted by
nonnegative definite matrices instead of nonnegative scalars, consensus and
clustering phenomena naturally exist. We examine algebraic and algebraic graph
conditions for achieving a consensus, and provide an algorithm for finding all
clusters of a given system. Finally, we illustrate two applications of the
proposed consensus algorithm in clustered consensus and in bearing-based
formation control.
| 0 | 0 | 1 | 0 | 0 | 0 |
Multi-armed Bandit Problems with Strategic Arms | We study a strategic version of the multi-armed bandit problem, where each
arm is an individual strategic agent and we, the principal, pull one arm each
round. When pulled, the arm receives some private reward $v_a$ and can choose
an amount $x_a$ to pass on to the principal (keeping $v_a-x_a$ for itself). All
non-pulled arms get reward $0$. Each strategic arm tries to maximize its own
utility over the course of $T$ rounds. Our goal is to design an algorithm for
the principal incentivizing these arms to pass on as much of their private
rewards as possible.
When private rewards are stochastically drawn each round ($v_a^t \leftarrow
D_a$), we show that:
- Algorithms that perform well in the classic adversarial multi-armed bandit
setting necessarily perform poorly: For all algorithms that guarantee low
regret in an adversarial setting, there exist distributions $D_1,\ldots,D_k$
and an approximate Nash equilibrium for the arms where the principal receives
reward $o(T)$.
- Still, there exists an algorithm for the principal that induces a game
among the arms where each arm has a dominant strategy. When each arm plays its
dominant strategy, the principal sees expected reward $\mu'T - o(T)$, where
$\mu'$ is the second-largest of the means $\mathbb{E}[D_{a}]$. This algorithm
maintains its guarantee if the arms are non-strategic ($x_a = v_a$), and also
if there is a mix of strategic and non-strategic arms.
| 1 | 0 | 0 | 1 | 0 | 0 |
Emerging Topics in Assistive Reading Technology: From Presentation to Content Accessibility | With the recent focus in the accessibility field, researchers from academia
and industry have been very active in developing innovative techniques and
tools for assistive technology. Especially with handheld devices getting ever
powerful and being able to recognize the user's voice, screen magnification for
individuals with low-vision, and eye tracking devices used in studies with
individuals with physical and intellectual disabilities, the science field is
quickly adapting and creating conclusions as well as products to help. In this
paper, we will focus on new technology and tools to help make reading
easier--including reformatting document presentation (for people with physical
vision impairments) and text simplification to make information itself easier
to interpret (for people with intellectual disabilities). A real-world case
study is reported based on our experience to make documents more accessible.
| 1 | 0 | 0 | 0 | 0 | 0 |
Supervised learning with quantum enhanced feature spaces | Machine learning and quantum computing are two technologies each with the
potential for altering how computation is performed to address previously
untenable problems. Kernel methods for machine learning are ubiquitous for
pattern recognition, with support vector machines (SVMs) being the most
well-known method for classification problems. However, there are limitations
to the successful solution to such problems when the feature space becomes
large, and the kernel functions become computationally expensive to estimate. A
core element to computational speed-ups afforded by quantum algorithms is the
exploitation of an exponentially large quantum state space through controllable
entanglement and interference. Here, we propose and experimentally implement
two novel methods on a superconducting processor. Both methods represent the
feature space of a classification problem by a quantum state, taking advantage
of the large dimensionality of quantum Hilbert space to obtain an enhanced
solution. One method, the quantum variational classifier builds on [1,2] and
operates through using a variational quantum circuit to classify a training set
in direct analogy to conventional SVMs. In the second, a quantum kernel
estimator, we estimate the kernel function and optimize the classifier
directly. The two methods present a new class of tools for exploring the
applications of noisy intermediate scale quantum computers [3] to machine
learning.
| 0 | 0 | 0 | 1 | 0 | 0 |
On The Limitation of Some Fully Observable Multiple Session Resilient Shoulder Surfing Defense Mechanisms | Using password based authentication technique, a system maintains the login
credentials (username, password) of the users in a password file. Once the
password file is compromised, an adversary obtains both the login credentials.
With the advancement of technology, even if a password is maintained in hashed
format, then also the adversary can invert the hashed password to get the
original one. To mitigate this threat, most of the systems nowadays store some
system generated fake passwords (also known as honeywords) along with the
original password of a user. This type of setup confuses an adversary while
selecting the original password. If the adversary chooses any of these
honeywords and submits that as a login credential, then system detects the
attack. A large number of significant work have been done on designing
methodologies (identified as $\text{M}^{\text{DS}}_{\text{OA}}$) that can
protect password against observation or, shoulder surfing attack. Under this
attack scenario, an adversary observes (or records) the login information
entered by a user and later uses those credentials to impersonate the genuine
user. In this paper, we have shown that because of their design principle, a
large subset of $\text{M}^{\text{DS}}_{\text{OA}}$ (identified as
$\text{M}^{\text{FODS}}_{\text{SOA}}$) cannot afford to store honeywords in
password file. Thus these methods, belonging to
$\text{M}^{\text{FODS}}_{\text{SOA}}$, are unable to provide any kind of
security once password file gets compromised. Through our contribution in this
paper, by still using the concept of honeywords, we have proposed few generic
principles to mask the original password of
$\text{M}^{\text{FODS}}_{\text{SOA}}$ category methods. We also consider few
well-established methods like S3PAS, CHC, PAS and COP belonging to
$\text{M}^{\text{FODS}}_{\text{SOA}}$, to show that proposed idea is
implementable in practice.
| 1 | 0 | 0 | 0 | 0 | 0 |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.