Searcharxiv⌕ Search

arXiv subjects

Fei Qi

Publications and source records attributed to Fei Qi.

27 records · Page 2Linked to original sources

Calibrated Intervention and Containment of the COVID-19 Pandemic

Within a short period of time, COVID-19 grew into a world-wide pandemic. Transmission by pre-symptomatic and asymptomatic viral carriers rendered intervention and containment of the disease extremely challenging. Based on reported infection case studies, we construct an epidemiological model that focuses on transmission around the symptom onset. The model is calibrated against incubation period and pairwise transmission statistics during the initial outbreaks of the pandemic outside Wuhan with minimal non-pharmaceutical interventions. Mathematical treatment of the model yields explicit expressions for the size of latent and pre-symptomatic subpopulations during the exponential growth phase, with the local epidemic growth rate as input. We then explore reduction of the basic reproduction number R_0 through specific disease control measures such as contact tracing, testing, social distancing, wearing masks and sheltering in place. When these measures are implemented in combination, their effects on R_0 multiply. We also compare our model behaviour to the first wave of the COVID-19 spreading in various affected regions and highlight generic and less generic features of the pandemic development.

q-bio.PE↗

The first cohomology, derivations and the reductivity of a (meromorphic open-string) vertex algebra

We give a criterion for the complete reducibility of modules satisfying a composability condition for a meromorphic open-string vertex algebra $V$ using the first cohomology of the algebra. For a $V$-bimodule $M$, let $\hat{H}^{1}_{\infty}(V, M)$ be the first cohomology of $V$ with the coefficients in $M$. Let $\hat{Z}^{1}_{\infty}(V, M)$ be the subspace of $\hat{H}^{1}_{\infty}(V, M)$ canonically isomorphic to the space of derivations obtained from the zero mode of the right vertex operators of weight $1$ elements such that the difference between the skew-symmetric opposite action of the left action and the right action on these elements are Laurent polynomials in the variable. If $\hat{H}^{1}_{\infty}(V, M)= \hat{Z}^{1}_{\infty}(V, M)$ for every $\Z$-graded $V$-bimodule $M$, then every left $V$-module satisfying a composability condition is completely reducible. In particular, since a lower-bounded $\Z$-graded vertex algebra $V$ is a special meromorphic open-string vertex algebra and left $V$-modules are in fact what has been called generalized $V$-modules with lower-bounded weights (or lower-bounded generalized $V$-modules), this result provides a cohomological criterion for the complete reducibility of lower-bounded generalized modules for such a vertex algebra. We conjecture that the converse of the main theorem above is also true. We also prove that when a grading-restricted vertex algebra $V$ contains a subalgebra satisfying some familiar conditions, the composability condition for grading-restricted generalized $V$-modules always holds and we need $\hat{H}^{1}_{\infty}(V, M)= \hat{Z}^{1}_{\infty}(V, M)$ only for every $\Z$-graded $V$-bimodule $M$ generated by a grading-restricted subspace in our complete reducibility theorem.

math.QA↗

A Single-Shot Arbitrarily-Shaped Text Detector based on Context Attended Multi-Task Learning

Detecting scene text of arbitrary shapes has been a challenging task over the past years. In this paper, we propose a novel segmentation-based text detector, namely SAST, which employs a context attended multi-task learning framework based on a Fully Convolutional Network (FCN) to learn various geometric properties for the reconstruction of polygonal representation of text regions. Taking sequential characteristics of text into consideration, a Context Attention Block is introduced to capture long-range dependencies of pixel information to obtain a more reliable segmentation. In post-processing, a Point-to-Quad assignment method is proposed to cluster pixels into text instances by integrating both high-level object knowledge and low-level pixel information in a single shot. Moreover, the polygonal representation of arbitrarily-shaped text can be extracted with the proposed geometric properties much more effectively. Experiments on several benchmarks, including ICDAR2015, ICDAR2017-MLT, SCUT-CTW1500, and Total-Text, demonstrate that SAST achieves better or comparable performance in terms of accuracy. Furthermore, the proposed algorithm runs at 27.63 FPS on SCUT-CTW1500 with a Hmean of 81.0% on a single NVIDIA Titan Xp graphics card, surpassing most of the existing segmentation-based methods.

cs.CV↗

On the cohomology of meromorphic open-string vertex algebras

This paper generalizes Huang's cohomology theory of grading-restricted vertex algebras to meromorphic open-string vertex algebras (MOSVAs hereafter), which are noncommutative generalizations of grading-restricted vertex algebras introduced by Huang. The vertex operators for these algebras satisfy associativity but do not necessarily satisfy the commutativity. Moreover, the MOSVA and its bimodules considered in this paper do not necessarily have finite-dimensional homogeneous subspaces, though we do require that they have lower-bounded gradings. The construction and results in this paper will be used in a joint paper by Huang and the author to give a cohomological criterion of the reductivity for modules for grading-restricted vertex algebras

math.QA↗

On modules for meromorphic open-string vertex algebras

We study representations of the meromorphic open-string vertex algebra (MOSVAs hereafter) defined in [H3], a noncommutative generalization of vertex (operator) algebra. We start by recalling the definition of a MOSVA $V$ and left $V$-modules in [H3]. Then we define right $V$-modules and $V$-bimodules that reflect the noncommutative nature of $V$. When $V$ satisfies a condition on the order of poles of the correlation function (which we call pole-order condition), we prove that the rationality of products of two vertex operators implies the rationality of products of any numbers of vertex operators. Also, the rationality of iterates of any numbers of vertex operators is established, and is used to construct the opposite MOSVA $V^{op}$ of $V$. It is proved here that right (resp. left) $V$-modules are equivalent to left (resp. right) $V^{op}$-modules. Using this equivalence, we prove that if $V$ and a grading-restricted left $V$-module $W$ is endowed with a Möbius structure, then the graded dual $W'$ of $W$ is a right $V$-module. This proof is the only place in this paper that needs the grading-restriction condition. Also, this result is generalized to not-grading-restricted modules under a strong pole-order condition that is satisfied by all existing examples of MOSVAs and modules.

math.QA↗

DarwinML: A Graph-based Evolutionary Algorithm for Automated Machine Learning

As an emerging field, Automated Machine Learning (AutoML) aims to reduce or eliminate manual operations that require expertise in machine learning. In this paper, a graph-based architecture is employed to represent flexible combinations of ML models, which provides a large searching space compared to tree-based and stacking-based architectures. Based on this, an evolutionary algorithm is proposed to search for the best architecture, where the mutation and heredity operators are the key for architecture evolution. With Bayesian hyper-parameter optimization, the proposed approach can automate the workflow of machine learning. On the PMLB dataset, the proposed approach shows the state-of-the-art performance compared with TPOT, Autostacker, and auto-sklearn. Some of the optimized models are with complex structures which are difficult to obtain in manual design.

cs.NE↗

Interlocking mechanism between molecular gears attached to surfaces

While molecular machines play an increasingly significant role in nanoscience research and applications, there remains a shortage of investigations and understanding of the molecular gear (cogwheel), which is an indispensable and fundamental component to drive a larger correlated molecular machine system. Employing ab initio calculations, we investigate model systems consisting of molecules adsorbed on metal or graphene surfaces, ranging from very simple triple-arm gears such as PF3 and NH3 to larger multi-arm gears based on carbon rings. We explore in detail the transmission of slow rotational motion from one gear to the next by these relatively simple molecules, so as to isolate and reveal the mechanisms of the relevant intermolecular interactions. Several characteristics of molecular gears are discussed, in particular the flexibility of the arms and the slipping and skipping between interlocking arms of adjacent gears, which differ from familiar macroscopic rigid gears. The underlying theoretical concepts suggest strongly that other analogous structures may also exhibit similar behavior which may inspire future exploration in designing large correlated molecular machines.

physics.chem-ph↗

Intramolecular Torque, an Indicator of the Internal Rotation Direction of Rotor Molecules and Similar Systems

Torque is ubiquitous in many molecular systems, including collisions, chemical reactions, vibrations, electronic excitations and especially rotor molecules. We present a straightforward theoretical method based on forces acting on atoms and obtained from atomistic quantum mechanics calculations, to quickly and qualitatively determine whether a molecule or sub-unit thereof has a tendency to rotation and, if so, around which axis and in which sense: clockwise or counterclockwise. The method also indicates which atoms, if any, are predominant in causing the rotation. Our computational approach can in general efficiently provide insights into the rotational ability of many molecules and help to theoretically screen or modify them in advance of experiments or before analyzing their rotational behavior in more detail with more extensive computations guided by the results from the torque approach. As an example, we demonstrate the effectiveness of the approach using a specific light-driven molecular rotary motor which was successfully synthesized and analyzed in prior experiments and simulations.

physics.chem-ph↗

A motivated proof of the Göllnitz-Gordon-Andrews identities

We present what we call a "motivated proof" of the Göllnitz-Gordon-Andrews identities. A similar motivated proof of the Rogers-Ramanujan identities was previously given by G. E. Andrews and R. J. Baxter, and was subsequently generalized to Gordon's identities by J. Lepowsky and M. Zhu. We anticipate that the present proof of the Göllnitz-Gordon-Andrews identities will illuminate certain twisted vertex-algebraic constructions.

math.CO↗