Searcharxiv⌕ Search

arXiv subjects

Yiwei Zhang

Publications and source records attributed to Yiwei Zhang.

At least 109 records · Page 6Linked to original sources
arXiv2018

Modelling brain-wide neuronal morphology via rooted Cayley trees

Neuronal morphology is an essential element for brain activity and function. We take advantage of current availability of brain-wide neuron digital reconstructions of the Pyramidal cells from a mouse brain, and analyze several emergent features of brain-wide neuronal morphology. We observe that axonal trees are self-affine while dendritic trees are self-similar. We also show that tree size appear to be random, independent of the number of dendrites within single neurons. Moreover, we consider inhomogeneous branching model which stochastically generates rooted 3-Cayley trees for the brain-wide neuron topology. Based on estimated order-dependent branching probability from actual axonal and dendritic trees, our inhomogeneous model quantitatively captures a number of topological features including size and shape of both axons and dendrites. This sheds lights on a universal mechanism behind the topological formation of brain-wide axonal and dendritic trees.

math.DS↗
arXiv2018

A remark on the central model method for the weak Palis conjecture of higher dimensional singular flows

For a generic vector field robustly without horseshoes, and an aperiodic chain recurrent class with singularities whose saddle values have different signs, the extended rescaled Poincaré map is associated with a central model. We estimate such central model and show it must have chain recurrent central segments over the singularities. This obstructs the application of central model to create horseshoes, and indicates that, differing from $C^1$ diffeomorphisms, solo using central model method is insufficient as a strategy to prove weak Palis conjecture for higher dimensional ($\geq 4$) singular flows. Our computation is actually based on simplified way of addressing blowup construction. As a byproduct, we are applicable to directly compute the extended rescaled Poincaré map upto second order derivatives, which we believe has its independent interests.

math.DS↗
arXiv2018

Reinforcement Mechanism Design for e-commerce

We study the problem of allocating impressions to sellers in e-commerce websites, such as Amazon, eBay or Taobao, aiming to maximize the total revenue generated by the platform. We employ a general framework of reinforcement mechanism design, which uses deep reinforcement learning to design efficient algorithms, taking the strategic behaviour of the sellers into account. Specifically, we model the impression allocation problem as a Markov decision process, where the states encode the history of impressions, prices, transactions and generated revenue and the actions are the possible impression allocations in each round. To tackle the problem of continuity and high-dimensionality of states and actions, we adopt the ideas of the DDPG algorithm to design an actor-critic policy gradient algorithm which takes advantage of the problem domain in order to achieve convergence and stability. We evaluate our proposed algorithm, coined IA(GRU), by comparing it against DDPG, as well as several natural heuristics, under different rationality models for the sellers - we assume that sellers follow well-known no-regret type strategies which may vary in their degree of sophistication. We find that IA(GRU) outperforms all algorithms in terms of the total revenue.

cs.MA↗
arXiv2017

Private Information Retrieval from MDS Coded Databases with Colluding Servers under Several Variant Models

Private information retrieval (PIR) gets renewed attentions due to its information-theoretic reformulation and its application in distributed storage system (DSS). The general PIR model considers a coded database containing $N$ servers storing $M$ files. Each file is stored independently via the same arbitrary $(N,K)$-MDS code. A user wants to retrieve a specific file from the database privately against an arbitrary set of $T$ colluding servers. A key problem is to analyze the PIR capacity, defined as the maximal number of bits privately retrieved per one downloaded bit. Several extensions for the general model appear by bringing in various additional constraints. In this paper, we propose a general PIR scheme for several variant PIR models including: PIR with robust servers, PIR with Byzantine servers, the multi-file PIR model and PIR with arbitrary collusion patterns.

cs.IT↗
arXiv2017

Simulations of Fractal Star Cluster Formation: I. New Insights for Measuring Mass Segregation of Star Clusters with SubStructure

We compare the existent methods including the minimum spanning tree based method and the local stellar density based method, in measuring mass segregation of star clusters. We find that the minimum spanning tree method reflects more the compactness, which represents the global spatial distribution of massive stars, while the local stellar density method reflects more the crowdedness, which provides the local gravitational potential information. It is suggested to measure the local and the global mass segregation simultaneously. We also develop a hybrid method that takes both aspects into account. This hybrid method balances the local and the global mass segregation in the sense that the predominant one is either caused either by dynamical evolution or purely accidental, especially when such information is unknown a priori. In addition, we test our prescriptions with numerical models and show the impact of binaries in estimating the mass segregation value. As an application, we use these methods on the Orion Nebula Cluster (ONC) observations and the Taurus cluster. We find that the ONC is significantly mass segregated down to the 20th most massive stars. In contrast, the massive stars of the Taurus cluster are sparsely distributed in many different subclusters, showing a low degree of compactness. The massive stars of Taurus are also found to be distributed in the high-density region of the subclusters, showing significant mass segregation at subcluster scales. Meanwhile, we also apply these methods to discuss the possible mechanisms of the dynamical evolution of the simulated substructured star clusters.

astro-ph.GA↗
arXiv2017

A general private information retrieval scheme for MDS coded databases with colluding servers

The problem of private information retrieval gets renewed attentions in recent years due to its information-theoretic reformulation and applications in distributed storage systems. PIR capacity is the maximal number of bits privately retrieved per one bit of downloaded bit. The capacity has been fully solved for some degenerating cases. For a general case where the database is both coded and colluded, the exact capacity remains unknown. We build a general private information retrieval scheme for MDS coded databases with colluding servers. Our scheme achieves the rate $(1+R+R^2+\cdots+R^{M-1})$, where $R=1-\frac{{{N-T}\choose K}}{N\choose K}$. Compared to existing PIR schemes, our scheme performs better for a certain range of parameters and is suitable for any underlying MDS code used in the distributed storage system.

cs.IT↗
arXiv2016

New bounds of permutation codes under Hamming metric and Kendall's $τ$-metric

Permutation codes are widely studied objects due to their numerous applications in various areas, such as power line communications, block ciphers, and the rank modulation scheme for flash memories. Several kinds of metrics are considered for permutation codes according to their specific applications. This paper concerns some improvements on the bounds of permutation codes under Hamming metric and Kendall's $τ$-metric respectively, using mainly a graph coloring approach. Specifically, under Hamming metric, we improve the Gilbert-Varshamov bound asymptotically by a factor $n$, when the minimum Hamming distance $d$ is fixed and the code length $n$ goes to infinity. Under Kendall's $τ$-metric, we narrow the gap between the known lower bounds and upper bounds. Besides, we also obtain some sporadic results under Kendall's $τ$-metric for small parameters.

cs.IT↗
arXiv2016

On private information retrieval array codes

Given a database, the private information retrieval (PIR) protocol allows a user to make queries to several servers and retrieve a certain item of the database via the feedbacks, without revealing the privacy of the specific item to any single server. Classical models of PIR protocols require that each server stores a whole copy of the database. Recently new PIR models are proposed with coding techniques arising from distributed storage system. In these new models each server only stores a fraction $1/s$ of the whole database, where $s>1$ is a given rational number. PIR array codes are recently proposed by Fazeli, Vardy and Yaakobi to characterize the new models. Consider a PIR array code with $m$ servers and the $k$-PIR property (which indicates that these $m$ servers may emulate any efficient $k$-PIR protocol). The central problem is to design PIR array codes with optimal rate $k/m$. Our contribution to this problem is three-fold. First, for the case $1 2$, we derive a new upper bound on the rate of a PIR array code. Finally, for the case $s>2$, we analyze a new construction by Blackburn and Etzion and show that its rate is better than all the other existing constructions.

cs.IT↗
arXiv2016

Centralized coded caching schemes: A hypergraph theoretical approach

The centralized coded caching scheme is a technique proposed by Maddah-Ali and Niesen as a solution to reduce the network burden in peak times in a wireless system. Later Yan et al. reformulated the problem as designing a corresponding placement delivery array, and proposed two new schemes from this perspective. These schemes above significantly reduce the transmission rate $R$, compared with the uncoded caching scheme. However, to implement the new schemes, each file should be cut into $F$ pieces, where $F$ increases exponentially with the number of users $K$. Such constraint is obviously infeasible in the practical setting, especially when $K$ is large. Thus it is desirable to design caching schemes with constant rate $R$ (independent of $K$) as well as small $F$. In this paper we view the centralized coded caching problem in a hypergraph perspective and show that designing a feasible placement delivery array is equivalent to constructing a linear and (6, 3)-free 3-uniform 3-partite hypergraph. Several new results and constructions arise from our novel point of view. First, by using the famous (6, 3)-theorem in extremal combinatorics, we show that constant rate caching schemes with $F$ growing linearly with $K$ do not exist. Second, we present two infinite classes of centralized coded caching schemes, which include the schemes of Ali-Niesen and Yan et al. as special cases, respectively. Moreover, our constructions show that constant rate caching schemes with $F$ growing sub-exponentially with $K$ do exist.

cs.IT↗
arXiv2016

New Constructions of MDS Symbol-Pair Codes

Motivated by the application of high-density data storage technologies, symbol-pair codes are proposed to protect against pair-errors in symbol-pair channels, whose outputs are overlapping pairs of symbols. The research of symbol-pair codes with the largest minimum pair-distance is interesting since such codes have the best possible error-correcting capability. A symbol-pair code attaining the maximal minimum pair-distance is called a maximum distance separable (MDS) symbol-pair code. In this paper, we focus on constructing linear MDS symbol-pair codes over the finite field $\mathbb{F}_{q}$. We show that a linear MDS symbol-pair code over $\mathbb{F}_{q}$ with pair-distance $5$ exists if and only if the length $n$ ranges from $5$ to $q^2+q+1$. As for codes with pair-distance $6$, length ranging from $6$ to $q^{2}+1$, we construct linear MDS symbol-pair codes by using a configuration called ovoid in projective geometry. With the help of elliptic curves, we present a construction of linear MDS symbol-pair codes for any pair-distance $d+2$ with length $n$ satisfying $7\le d+2\leq n\le q+\lfloor 2\sqrt{q}\rfloor+δ(q)-3$, where $δ(q)=0$ or $1$.

cs.IT↗
arXiv2016

Invertible binary matrix with maximum number of $2$-by-$2$ invertible submatrices

The problem is related to all-or-nothing transforms (AONT) suggested by Rivest as a preprocessing for encrypting data with a block cipher. Since then there have been various applications of AONTs in cryptography and security. D'Arco, Esfahani and Stinson posed the problem on the constructions of binary matrices for which the desired properties of an AONT hold with the maximum probability. That is, for given integers $t\le s$, what is the maximum number of $t$-by-$t$ invertible submatrices in a binary matrix of order $s$? For the case $t=2$, let $R_2(s)$ denote the maximal proportion of 2-by-2 invertible submatrices. D'Arco, Esfahani and Stinson conjectured that the limit is between 0.492 and 0.625. We completely solve the case $t=2$ by showing that $\lim_{s\rightarrow\infty}R_2(s)=0.5$.

cs.IT↗
arXiv2015

On the mixing properties of piecewise expanding maps under composition with permutations

We consider the effect on the mixing properties of a piecewise smooth interval map $f$ when its domain is divided into $N$ equal subintervals and $f$ is composed with a permutation of these. The case of the stretch-and-fold map $f(x)=mx \bmod 1$ for integers $m \geq 2$ is examined in detail. We give a combinatorial description of those permutations $σ$ for which $σ\circ f$ is still (topologically) mixing, and show that the proportion of such permutations tends to $1$ as $N \to \infty$. We then investigate the mixing rate of $σ\circ f$ (as measured by the modulus of the second largest eigenvalue of the transfer operator). In contrast to the situation for continuous time diffusive systems, we show that composition with a permutation cannot improve the mixing rate of $f$, but typically makes it worse. Under some mild assumptions on $m$ and $N$, we obtain a precise value for the worst mixing rate as $σ$ ranges through all permutations; this can be made arbitrarily close to $1$ as $N \to \infty$ (with $m$ fixed). We illustrate the geometric distribution of the second largest eigenvalues in the complex plane for small $m$ and $N$, and propose a conjecture concerning their location in general. Finally, we give examples of other interval maps $f$ for which composition with permutations produces different behaviour than that obtained from the stretch-and-fold map.

math.DS↗
arXiv2015

Ergodic optimization of prevalent super-continuous functions

Given a dynamical system, we say that a performance function has property P if its time averages along orbits are maximized at a periodic orbit. It is conjectured by several authors that for sufficiently hyperbolic dynamical systems, property P should be typical among sufficiently regular performance functions. In this paper we address this problem using a probabilistic notion of typicality that is suitable to infinite dimension: the concept of prevalence as introduced by Hunt, Sauer, and Yorke. For the one-sided shift on two symbols, we prove that property P is prevalent in spaces of functions with a strong modulus of regularity. Our proof uses Haar wavelets to approximate the ergodic optimization problem by a finite-dimensional one, which can be conveniently restated as a maximum cycle mean problem on a de Bruijin graph.

math.DS↗
arXiv2015

Tree pressure for hyperbolic and non-exceptional upper semi-continuous potentials

In this note, we investigate the tree pressure for multi-modal interval maps with a certain class of hyperbolic and non-exceptional upper semi-continuous functions. In particular, we obtain a generalized version of Corollary 2.2 in the paper \cite{LRL14} by Li and Rivera-Letelier. This property will be used to prove the existence of a conformal measure for the geometric potential in the negative spectrum.

math.DS↗
arXiv2015

Thermodynamic formalism of interval maps for upper semi-continuous potentials: Makarov-Smirnov's formalism

In this paper, we study the thermodynamic formalism of interval maps $f$ with sufficient regularity, for a sub class $\mathcal{U}$ composed of upper semi-continuous potentials which includes both Hölder and geometric potentials. We show that for a given $u\in \mathcal{U}$ and negative values of $t$, the pressure function $P(f,-tu)$ can be calculated in terms of the corresponding hidden pressure function $\widetilde{P}(f,-tu)$. Determination of the values $t\in(-\infty,0)$ at which $P(f,-tu)\neq \widetilde{P}(f,-tu)$ is also characterized explicitly. When restricting to the Hölder continuous potentials, our result recovers Theorem B in [Li \& Rivera-Letelier 2013] for maps with non-flat critical points. While restricting to the geometric potentials, we develop a real version of Makarnov-Smirnov's formalism, in parallel to the complex version shown in [Makarnov \& Smirnov 2000, Theo A,B]. Moreover, our results also provide a new and simpler proof (using [Ruelle 1992, Coro6.3]) of the original Makarnov-Smirnov's formalism in the complex setting, under an additional assumption about non-exceptionality, i.e., [Makarnov \& Smirnov 2000, Theo 3.1].

math.DS↗
arXiv2015

On the mixing properties of piecewise expanding maps under composition with permutations, II: Maps of non-constant orientation

For an integer $m \geq 2$, let $\mathcal{P}_m$ be the partition of the unit interval $I$ into $m$ equal subintervals, and let $\mathcal{F}_m$ be the class of piecewise linear maps on $I$ with constant slope $\pm m$ on each element of $\mathcal{P}_m$. We investigate the effect on mixing properties when $f \in \mathcal{F}_m$ is composed with the interval exchange map given by a permutation $σ\in S_N$ interchanging the $N$ subintervals of $\mathcal{P}_N$. This extends the work in a previous paper [N.P. Byott, M. Holland and Y. Zhang, DCDS, {\bf 33}, (2013) 3365--3390], where we considered only the "stretch-and-fold" map $f_{sf}(x)=mx \bmod 1$.

math.DS↗
arXiv2015

Snake-in-the-Box Codes for Rank Modulation under Kendall's $τ$-Metric

For a Gray code in the scheme of rank modulation for flash memories, the codewords are permutations and two consecutive codewords are obtained using a push-to-the-top operation. We consider snake-in-the-box codes under Kendall's $τ$-metric, which is a Gray code capable of detecting one Kendall's $τ$-error. We answer two open problems posed by Horovitz and Etzion. Firstly, we prove the validity of a construction given by them, resulting in a snake of size $M_{2n+1}=\frac{(2n+1)!}{2}-2n+1$. Secondly, we come up with a different construction aiming at a longer snake of size $M_{2n+1}=\frac{(2n+1)!}{2}-2n+3$. The construction is applied successfully to $S_7$.

cs.IT↗