ZWZhengdao Wang

Zhengdao Wang

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Zhengdao Wang has a total of 48 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Performance Analysis of Clustered Wireless-Powered Ad Hoc Networks via β-Ginibre Point Processes

    by Chao Zhang, Fangzhou Yu, Zhengdao Wang, Gangming Lyu

    We consider a wireless-powered ad hoc network with clustered Power Beacons (PBs). The wireless-powered transmitters (WP-Txs) only activate stochastically deployed PBs within the circles centered at their locations. The PBs transmit radio frequency (RF) signals, which are harvested by the WP-Txs to transmit information to their target receivers. The nodes are modeled as following $\beta $ -Ginibre Point Processes ( $\beta $ -GPP), which incorporate repulsion among PBs or active WP-Txs. Lacking analytical mathematical tools to describe the distribution of clustered PBs, we establish lower and upper bounds on the energy outage probability, as well as an approximated expression of it. We then derive the information outage probability, an approximate expression of it by approximating the reduced Palm distribution of $\beta $ -GPP as an inhomogeneous Poisson point process (IPPP), and an analytical approximate expression of it by further ignoring the small-scale fading of interfering WP-Tx links. Finally, the transmission capacity is obtained based on the overall outage probability. The simulation results show that the derived theoretical expressions accurately reflect the performance of the wireless-powered ad hoc networks and the performance gap introduced by approximations is negligible. Transmission capacity optimization based on the derived expressions is also possible and demonstrated.


  • PINE: Universal Deep Embedding for Graph Nodes via Partial Permutation Invariant Set Functions

    by Shupeng Gui, Xiangliang Zhang, Pan Zhong, Shuang Qiu, Mingrui Wu, Jieping Ye, Zhengdao Wang, Ji Liu

    Graph node embedding aims at learning a vector representation for all nodes given a graph. It is a central problem in many machine learning tasks (e.g., node classification, recommendation, community detection). The key problem in graph node embedding lies in how to define the dependence to neighbors Existing approaches specify (either explicitly or implicitly) certain dependencies on neighbors, which may lead to loss of subtle but important structural information within the graph and other dependence among neighbors. This intrigues us to ask the question: can we design a model to give the adaptive flexibility of dependencies to each node's neighborhood. In this paper, we propose a novel graph node embedding method (named PINE) via a novel notion of partial permutation invariant set function, to capture any possible dependence. Our method 1) can learn an arbitrary form of the representation function from the neighborhood, without losing any potential dependence structures, and 2) is applicable to both homogeneous and heterogeneous graph embedding, the latter of which is challenged by the diversity of node types. Furthermore, we provide theoretical guarantee for the representation capability of our method for general homogeneous and heterogeneous graphs.


  • Open Access

    Reinforcement Learning Architectures: SAC, TAC, and ESAC

    by Ala'Eddin Masadeh, Zhengdao Wang, Ahmed E. Kamal

    The trend is to implement intelligent agents capable of analyzing available information and utilize it efficiently. This work presents a number of reinforcement learning (RL) architectures; one of them is designed for intelligent agents. The proposed architectures are called selector-actor-critic (SAC), tuner-actor-critic (TAC), and estimator-selector-actor-critic (ESAC). These architectures are improved models of a well known architecture in RL called actor-critic (AC). In AC, an actor optimizes the used policy, while a critic estimates a value function and evaluate the optimized policy by the actor. SAC is an architecture equipped with an actor, a critic, and a selector. The selector determines the most promising action at the current state based on the last estimate from the critic. TAC consists of a tuner, a model-learner, an actor, and a critic. After receiving the approximated value of the current state-action pair from the critic and the learned model from the model-learner, the tuner uses the Bellman equation to tune the value of the current state-action pair. ESAC is proposed to implement intelligent agents based on two ideas, which are lookahead and intuition. Lookahead appears in estimating the values of the available actions at the next state, while the intuition appears in maximizing the probability of selecting the most promising action. The newly added elements are an underlying model learner, an estimator, and a selector. The model learner is used to approximate the underlying model. The estimator uses the approximated value function, the learned underlying model, and the Bellman equation to estimate the values of all actions at the next state. The selector is used to determine the most promising action at the next state, which will be used by the actor to optimize the used policy. Finally, the results show the superiority of ESAC compared with the other architectures.

  • Open Access

    Polarized Low-Density Parity-Check Codes on the BSC

    by Binbin Gao, Huarui Yin, Zhengdao Wang

    The connections between variable nodes and check nodes have a great influence on the performance of low-density parity-check (LDPC) codes. Inspired by the unique structure of polar code's generator matrix, we proposed a new method of constructing LDPC codes that achieves a polarization effect. The new code, named as polarized LDPC codes, is shown to achieve lower or no error floor in the binary symmetric channel (BSC)

  • Open Access

    PINE: Universal Deep Embedding for Graph Nodes via Partial Permutation Invariant Set Functions

    by Shupeng Gui, Xiangliang Zhang, Pan Zhong, Shuang Qiu, Mingrui Wu, Jieping Ye, Zhengdao Wang, Ji Liu

    Graph node embedding aims at learning a vector representation for all nodes given a graph. It is a central problem in many machine learning tasks (e.g., node classification, recommendation, community detection). The key problem in graph node embedding lies in how to define the dependence to neighbors. Existing approaches specify (either explicitly or implicitly) certain dependencies on neighbors, which may lead to loss of subtle but important structural information within the graph and other dependencies among neighbors. This intrigues us to ask the question: can we design a model to give the maximal flexibility of dependencies to each node's neighborhood. In this paper, we propose a novel graph node embedding (named PINE) via a novel notion of partial permutation invariant set function, to capture any possible dependence. Our method 1) can learn an arbitrary form of the representation function from the neighborhood, withour losing any potential dependence structures, and 2) is applicable to both homogeneous and heterogeneous graph embedding, the latter of which is challenged by the diversity of node types. Furthermore, we provide theoretical guarantee for the representation capability of our method for general homogeneous and heterogeneous graphs. Empirical evaluation results on benchmark data sets show that our proposed PINE method outperforms the state-of-the-art approaches on producing node vectors for various learning tasks of both homogeneous and heterogeneous graphs.

  • Spatial Transmitter Density Allocation for Frequency-Selective Wireless Ad Hoc Networks

    by Songtao Lu, Zhengdao Wang

    We consider a network of pairs of nodes that perform ad hoc simultaneous communications over frequency-selective channels. We assume that the whole frequency band is divided into a number of subbands, and each transmitter can only use one subband. Assuming that the network is geometrically infinite, we use the transmission capacity (TC) as a measure of network throughput. We consider the problem of allocating nodes to the subbands so that the total TC is maximized, under the constraint of a fixed total spatial node density. The optimization problem turns out to be nonconvex. We investigate the detailed structure of the functions involved in the optimization and identify a set of properties of the optimal transmitter density over the subbands. An iterative resource allocation scheme with low complexity is derived to obtain the global optimal solution of the TC maximization problem. The solution can be loosely interpreted as a water-filling solution for a nonconvex optimization problem. Based on numerical simulations, it is shown that the optimal solution obtained through the theoretical analysis is consistent with the one obtained through an exhaustive search, which reveals that the outage probability and total spatial transmitter density are the keys of determining the network TC.


  • Training Optimization and Performance of Single Cell Uplink System With Massive-Antennas Base Station

    by Songtao Lu, Zhengdao Wang

    We study the performance of uplink transmission in single-cell wireless systems, where all the transmitters have single antennas and the base station has a large number of antennas. We consider both maximum ratio combining and zero-forcing receivers and both small- and large-scale fading channels. We also characterize the achievable total degrees of freedom (DoF) of such a system without assuming channel state information at the receiver. The system DoF turns out to be the same as that of a single-user multiple-input multiple-output system. However, when the number of users is the same as the number of receive antennas, linear receivers are not sufficient for achieving the maximum total DoF. The amount of energy savings that are possible through the increased number of base-station antennas or increased coherence interval are quantified. Furthermore, the training period and training energy allocation under the average and peak power constraints are optimized jointly to maximize the achievable sum spectral efficiency (SE). The improvement on achievable SE provided by the training duration and energy optimization is verified through multiple numerical simulations.


  • On the Secret Key Capacity of Sibling Hidden Markov Models

    by Mohammad Reza Khalili Shoja, George T. Amariucai, Zhengdao Wang, Shuangqing Wei, Jing Deng

    Traditional approaches to secret key establishment based on common randomness have been based on certain restrictive assumptions, such as considering the available common randomness to consist of independent and identically distributed (i.i.d) repetitions of correlated random variables. Unfortunately, the i.i.d assumption does not generally reflect the conditions of real-life scenarios. For this reason, the current paper investigates the key-establishment potential of a more pragmatic model, in which all parties have access to imperfect information about a common source modeled as a Markov chain. Each party's information thus comes in the form of a hidden Markov model and, since the different parties share the same underlying Markov chain, we call the overall model a sibling hidden Markov model (SHMM). This paper studies upper and lower bounds on the secret key capacity for various types of SHMM. The difficulty of the problem emerges from its prohibitive computational cost. To address this obstacle, we represent the joint probability of the observations as the L 1 norm of a Markov random matrix, and use its convergence to a Lyapunov exponent.