arXiv subjects

Ji Liu

Publications and source records attributed to Ji Liu.

At least 307 records · Page 17Linked to original sources

arXiv2017

A Scalable Algorithm for Structure Identification of Complex Gene Regulatory Network from Temporal Expression Data

Motivation: Gene regulatory interactions are of fundamental importance to various biological functions and processes. However, only a few previous computational studies have claimed success in revealing genome-wide regulatory landscapes from temporal gene expression data, especially for complex eukaryotes like human. Moreover, recent work suggests that these methods still suffer from the curse of dimensionality if network size increases to 100 or higher. Result: We present a novel scalable algorithm for identifying genome-wide regulatory network structures. The highlight of our method is that its superior performance does not degenerate even for a network size on the order of $10^4$, and is thus readily applicable to large-scale complex networks. Such a breakthrough is achieved by considering both prior biological knowledge and multiple topological properties (i.e., sparsity and hub gene structure) of complex networks in the regularized formulation. We also illustrate the application of our algorithm in practice using the time-course expression data from an influenza infection study in respiratory epithelial cells. Availability and Implementation: The algorithm described in this article is implemented in MATLAB$^\circledR$. The source code is freely available from https://github.com/Hongyu-Miao/DMI.git. Contact: jliu@cs.rochester.edu; hongyu.miao@uth.tmc.edu Supplementary information: Supplementary data are available online.

q-bio.MN↗

arXiv2017

Exponential Convergence of a Distributed Algorithm for Solving Linear Algebraic Equations

In a recent paper, a distributed algorithm was proposed for solving linear algebraic equations of the form $Ax = b$ assuming that the equation has at least one solution. The equation is presumed to be solved by $m$ agents assuming that each agent knows a subset of the rows of the matrix $[A \; b]$, the current estimates of the equation's solution generated by each of its neighbors, and nothing more. Neighbor relationships are represented by a time-dependent directed graph $N(t)$ whose vertices correspond to agents and whose arcs characterize neighbor relationships. Sufficient conditions on $N(t)$ were derived under which the algorithm can cause all agents' estimates to converge exponentially fast to the same solution to $Ax = b$. These conditions were also shown to be necessary for exponential convergence, provided the data about $[A \; b]$ available to the agents is "non-redundant". The aim of this paper is to relax this "non-redundant" assumption. This is accomplished by establishing exponential convergence under conditions which are the weakest possible for the problem at hand; the conditions are based on a new notion of graph connectivity. An improved bound on the convergence rate is also derived.

math.OC↗

arXiv2016

Request-Based Gossiping without Deadlocks

By the distributed averaging problem is meant the problem of computing the average value of a set of numbers possessed by the agents in a distributed network using only communication between neighboring agents. Gossiping is a well-known approach to the problem which seeks to iteratively arrive at a solution by allowing each agent to interchange information with at most one neighbor at each iterative step. Crafting a gossiping protocol which accomplishes this is challenging because gossiping is an inherently collaborative process which can lead to deadlocks unless careful precautions are taken to ensure that it does not. Many gossiping protocols are request-based which means simply that a gossip between two agents will occur whenever one of the two agents accepts a request to gossip placed by the other. In this paper, we present three deterministic request-based protocols. We show by example that the first can deadlock. The second is guaranteed to avoid deadlocks and requires fewer transmissions per iteration than standard broadcast-based distributed averaging protocols by exploiting the idea of local ordering together with the notion of an agent's neighbor queue; the protocol requires the simplest queue updates, which provides an in-depth understanding of how local ordering and queue updates avoid deadlocks. It is shown that a third protocol which uses a slightly more complicated queue update rule can lead to significantly faster convergence; a worst case bound on convergence rate is provided.

math.OC↗

arXiv2016

Prognostics of Surgical Site Infections using Dynamic Health Data

Surgical Site Infection (SSI) is a national priority in healthcare research. Much research attention has been attracted to develop better SSI risk prediction models. However, most of the existing SSI risk prediction models are built on static risk factors such as comorbidities and operative factors. In this paper, we investigate the use of the dynamic wound data for SSI risk prediction. There have been emerging mobile health (mHealth) tools that can closely monitor the patients and generate continuous measurements of many wound-related variables and other evolving clinical variables. Since existing prediction models of SSI have quite limited capacity to utilize the evolving clinical data, we develop the corresponding solution to equip these mHealth tools with decision-making capabilities for SSI prediction with a seamless assembly of several machine learning models to tackle the analytic challenges arising from the spatial-temporal data. The basic idea is to exploit the low-rank property of the spatial-temporal data via the bilinear formulation, and further enhance it with automatic missing data imputation by the matrix completion technique. We derive efficient optimization algorithms to implement these models and demonstrate the superior performances of our new predictive model on a real-world dataset of SSI, compared to a range of state-of-the-art methods.

cs.LG↗

arXiv2016

Efficient Estimation of Compressible State-Space Models with Application to Calcium Signal Deconvolution

In this paper, we consider linear state-space models with compressible innovations and convergent transition matrices in order to model spatiotemporally sparse transient events. We perform parameter and state estimation using a dynamic compressed sensing framework and develop an efficient solution consisting of two nested Expectation-Maximization (EM) algorithms. Under suitable sparsity assumptions on the innovations, we prove recovery guarantees and derive confidence bounds for the state estimates. We provide simulation studies as well as application to spike deconvolution from calcium imaging data which verify our theoretical results and show significant improvement over existing algorithms.

stat.ML↗

arXiv2016

Accelerating Stochastic Composition Optimization

Consider the stochastic composition optimization problem where the objective is a composition of two expected-value functions. We propose a new stochastic first-order method, namely the accelerated stochastic compositional proximal gradient (ASC-PG) method, which updates based on queries to the sampling oracle using two different timescales. The ASC-PG is the first proximal gradient method for the stochastic composition problem that can deal with nonsmooth regularization penalty. We show that the ASC-PG exhibits faster convergence than the best known algorithms, and that it achieves the optimal sample-error complexity in several important special cases. We further demonstrate the application of ASC-PG to reinforcement learning and conduct numerical experiments.

math.OC↗

arXiv2016

Staleness-aware Async-SGD for Distributed Deep Learning

Deep neural networks have been shown to achieve state-of-the-art performance in several machine learning tasks. Stochastic Gradient Descent (SGD) is the preferred optimization algorithm for training these networks and asynchronous SGD (ASGD) has been widely adopted for accelerating the training of large-scale deep networks in a distributed computing environment. However, in practice it is quite challenging to tune the training hyperparameters (such as learning rate) when using ASGD so as achieve convergence and linear speedup, since the stability of the optimization algorithm is strongly influenced by the asynchronous nature of parameter updates. In this paper, we propose a variant of the ASGD algorithm in which the learning rate is modulated according to the gradient staleness and provide theoretical guarantees for convergence of this algorithm. Experimental verification is performed on commonly-used image classification benchmarks: CIFAR10 and Imagenet to demonstrate the superior effectiveness of the proposed approach, compared to SSGD (Synchronous SGD) and the conventional ASGD algorithm.

cs.LG↗

arXiv2016

On the Analysis of a Continuous-Time Bi-Virus Model

Motivated by the spread of opinions on different social networks, we study a distributed continuous-time bi-virus model for a system of groups of individuals. An in-depth stability analysis is performed for more general models than have been previously considered, for the healthy and epidemic states. In addition, we investigate sensitivity properties of some nontrivial equilibria and obtain an impossibility result for distributed feedback control.

math.OC↗

arXiv2016

On a Modified DeGroot-Friedkin Model of Opinion Dynamics

This paper studies the opinion dynamics that result when individuals consecutively discuss a sequence of issues. Specifically, we study how individuals' self-confidence levels evolve via a reflected appraisal mechanism. Motivated by the DeGroot-Friedkin model, we propose a Modified DeGroot-Friedkin model which allows individuals to update their self-confidence levels by only interacting with their neighbors and in particular, the modified model allows the update of self-confidence levels to take place in finite time without waiting for the opinion process to reach a consensus on any particular issue. We study properties of this Modified DeGroot-Friedkin model and compare the associated equilibria and stability with those of the original DeGroot-Friedkin model. Specifically, for the case when the interaction matrix is doubly stochastic, we show that for the modified model, the vector of individuals' self-confidence levels asymptotically converges to a unique nontrivial equilibrium which for each individual is equal to 1/n, where n is the number of individuals. This implies that eventually, individuals reach a democratic state.

math.OC↗

arXiv2015

Exponential Convergence of the Discrete-Time Altafini Model

This paper considers the discrete-time version of Altafini's model for opinion dynamics in which the interaction among a group of agents is described by a time-varying signed digraph. Prompted by an idea from [1], exponential convergence of the system is studied using a graphical approach. Necessary and sufficient conditions for exponential convergence with respect to each possible type of limit states are provided. Specifically, under the assumption of repeatedly jointly strong connectivity, it is shown that (1) a certain type of two-clustering will be reached exponentially fast for almost all initial conditions if, and only if, the sequence of signed digraphs is repeatedly jointly structurally balanced corresponding to that type of two-clustering; (2) the system will converge to zero exponentially fast for all initial conditions if, and only if, the sequence of signed digraphs is repeatedly jointly structurally unbalanced. An upper bound on the convergence rate is also provided.

math.OC↗

arXiv2015

Asynchronous Stochastic Coordinate Descent: Parallelism and Convergence Properties

We describe an asynchronous parallel stochastic proximal coordinate descent algorithm for minimizing a composite objective function, which consists of a smooth convex function plus a separable convex function. In contrast to previous analyses, our model of asynchronous computation accounts for the fact that components of the unknown vector may be written by some cores simultaneously with being read by others. Despite the complications arising from this possibility, the method achieves a linear convergence rate on functions that satisfy an optimal strong convexity property and a sublinear rate ($1/k$) on general convex functions. Near-linear speedup on a multicore system can be expected if the number of processors is $O(n^{1/4})$. We describe results from implementation on ten cores of a multicore processor.

math.OC↗

arXiv2015

The Teaching Dimension of Linear Learners

Teaching dimension is a learning theoretic quantity that specifies the minimum training set size to teach a target model to a learner. Previous studies on teaching dimension focused on version-space learners which maintain all hypotheses consistent with the training data, and cannot be applied to modern machine learners which select a specific hypothesis via optimization. This paper presents the first known teaching dimension for ridge regression, support vector machines, and logistic regression. We also exhibit optimal training sets that match these teaching dimensions. Our approach generalizes to other linear learners.

cs.LG↗

arXiv2015

Parametric cooling of a degenerate Fermi gas in an optical trap

We demonstrate a novel technique for cooling a degenerate Fermi gas in a crossed-beam optical dipole trap, where high-energy atoms can be selectively removed from the trap by modulating the stiffness of the trapping potential with anharmonic trapping frequencies. We measure the dependence of the cooling effect on the frequency and amplitude of the parametric modulations. It is found that the large anharmonicity along the axial trapping potential allows to generate a degenerate Fermi gas with anisotropic energy distribution, in which the cloud energy in the axial direction can be reduced to the ground state value.

cond-mat.quant-gas↗

arXiv2015

Distributed Evaluation and Convergence of Self-Appraisals in Social Networks

We consider in this paper a networked system of opinion dynamics in continuous time, where the agents are able to evaluate their self-appraisals in a distributed way. In the model we formulate, the underlying network topology is described by a rooted digraph. For each ordered pair of agents $(i,j)$, we assign a function of self-appraisal to agent $i$, which measures the level of importance of agent $i$ to agent $j$. Thus, by communicating only with her neighbors, each agent is able to calculate the difference between her level of importance to others and others' level of importance to her. The dynamical system of self-appraisals is then designed to drive these differences to zero. We show that for almost all initial conditions, the trajectory generated by this dynamical system asymptotically converges to an equilibrium point which is exponentially stable.

eess.SY↗

arXiv2015

Exclusive Sparsity Norm Minimization with Random Groups via Cone Projection

Many practical applications such as gene expression analysis, multi-task learning, image recognition, signal processing, and medical data analysis pursue a sparse solution for the feature selection purpose and particularly favor the nonzeros \emph{evenly} distributed in different groups. The exclusive sparsity norm has been widely used to serve to this purpose. However, it still lacks systematical studies for exclusive sparsity norm optimization. This paper offers two main contributions from the optimization perspective: 1) We provide several efficient algorithms to solve exclusive sparsity norm minimization with either smooth loss or hinge loss (non-smooth loss). All algorithms achieve the optimal convergence rate $O(1/k^2)$ ($k$ is the iteration number). To the best of our knowledge, this is the first time to guarantee such convergence rate for the general exclusive sparsity norm minimization; 2) When the group information is unavailable to define the exclusive sparsity norm, we propose to use the random grouping scheme to construct groups and prove that if the number of groups is appropriately chosen, the nonzeros (true features) would be grouped in the ideal way with high probability. Empirical studies validate the efficiency of proposed algorithms, and the effectiveness of random grouping scheme on the proposed exclusive SVM formulation.

stat.ML↗

arXiv2015

Products of Generalized Stochastic Sarymsakov Matrices

In the set of stochastic, indecomposable, aperiodic (SIA) matrices, the class of stochastic Sarymsakov matrices is the largest known subset (i) that is closed under matrix multiplication and (ii) the infinitely long left-product of the elements from a compact subset converges to a rank-one matrix. In this paper, we show that a larger subset with these two properties can be derived by generalizing the standard definition for Sarymsakov matrices. The generalization is achieved either by introducing an "SIA index", whose value is one for Sarymsakov matrices, and then looking at those stochastic matrices with larger SIA indices, or by considering matrices that are not even SIA. Besides constructing a larger set, we give sufficient conditions for generalized Sarymsakov matrices so that their products converge to rank-one matrices. The new insight gained through studying generalized Sarymsakov matrices and their products has led to a new understanding of the existing results on consensus algorithms and will be helpful for the design of network coordination algorithms.

math.OC↗

arXiv2015

A Stackelberg Game for Multi-Period Demand Response Management in the Smart Grid

This paper studies a multi-period demand response management problem in the smart grid where multiple utility companies compete among themselves. The user-utility interactions are modeled by a noncooperative game of a Stackelberg type where the interactions among the utility companies are captured through a Nash equilibrium. It is shown that this game has a unique Stackelberg equilibrium at which the utility companies set prices to maximize their revenues (within a Nash game) while the users respond accordingly to maximize their utilities subject to their budget constraints. Closed-form expressions are provided for the corresponding strategies of the users and the utility companies. It is shown that the multi- period scheme, compared with the single-period case, provides more incentives for the users to participate in the game. A necessary and sufficient condition on the minimum budget needed for a user to participate is provided.

math.OC↗

arXiv2015

Complete scaling makes differences along the critical isobar: Molecular dynamics simulations of Lennard-Jones fluid with finite-time scaling

We show that, along the critical isobar, the complete scaling results in a unique leading scaling qualitatively distinct to that arising from the simple and the revised scalings. This is verified by a complete-field finite-time scaling theory, which combines the complete scaling with finite-time scaling, and its application to the molecular dynamics simulations of the vapor-liquid critical point of a three-dimensional one-component Lennard-Jones fluid in an isobaric-isothermal ensemble with linear heating or cooling. Both the static and the dynamic critical exponents as well as the critical parameters can be estimated without \emph{a priori} knowledge of the universality class. The results agree with extant values and thus show the necessity of the complete scaling to the leading asymptotic behavior along the critical isobar even for the LJ fluid whose asymmetry is thought to be weak.

cond-mat.stat-mech↗