Information cascades represent traces of information propagation in latent unknown networks. The underlying information diffusion dynamics are captured from collected cascade data and used to infer online social network connectivity and structure which is useful for several applications such as influence maximization and rumor blocking. Our goal is to make the network inference process robust to missing cascade entries. As a first approach, we attempt to recover the missing elements and use the recovered cascades to solve the inference problem. The second approach consists in using expectation maximization algorithm to jointly recover missing entries and solve the network inference problem iteratively. Experiments are implemented on synthetic as well as large scale real social networks to validate and compare the proposed methods.

Learning and exploring the connectivity of unknown networks represent an important problem in practical applications of communication networks and social-media networks. Modeling large-scale networks as connected graphs is highly desirable to extract their connectivity information among nodes to visualize network topology, disseminate data, and improve routing efficiency. This paper investigates a simple measurement model in which a small subset of source nodes collect hop distance information from networked nodes in order to generate a virtual coordinate system (VCS) for networks of unknown topology. We establish the VCS to define logical distance among nodes based on principal component analysis and to determine connectivity relationship and effective routing methods. More importantly, we present a robust analytical algorithm to derive the VCS against practical issues of missing and corrupted measurements. We also develop a connectivity inference method which classifies nodes into layers based on the hop distances and derives partial information on network connectivity.

Learning network topology from partial knowledge of its connectivity is an important objective in practical scenarios of communication networks and social-media networks. Representing such networks as connected graphs, exploring and recovering connectivity information between network nodes can help visualize the network topology and improve network utility. This work considers the use of simple hop distance measurement obtained from a fraction of anchor/source nodes to reconstruct the node connectivity relationship for large scale networks of unknown connection topology. Our proposed approach consists of two steps. We first develop a tree-based search strategy to determine constraints on unknown network edges based on the hop count measurements. We then derive the logical distance between nodes based on principal component analysis (PCA) of the measurement matrix and propose a binary hypothesis test for each unknown edge. The proposed algorithm can effectively improve both the accuracy of connectivity detection and the successful delivery rate in data routing applications.

In multiple-input multiple-output radar systems, it is usually desirable to steer transmitted power in the region-of-interest. To do this, conventional methods optimize the waveform covariance matrix,R, for the desired beampattern, which is then used to generate actual transmitted waveforms. In this paper, we provide a low complexity closed-form solution to design covariance matrix for the given planar beampattern using the planar array, which is then used to derive a novel closedform algorithm to directly design the finite-alphabet constantenvelope waveforms. The proposed algorithm exploits the two-dimensional fast-Fourier-transform. The performance of our proposed algorithm is compared with the existing methods that are based on semi-definite quadratic programming with the advantage of a considerably reduced complexity.

We consider wireless sensor networks (WSNs) used for distributed estimation of unknown parameters. Due to the limited bandwidth, sensor nodes quantize their noisy observations before transmission to a fusion center (FC) for the estimation process. In this letter, the correlation between observations is exploited to reduce the mean-square error (MSE) of the distributed estimation. Specifically, sensor nodes generate local predictions of their observations and then transmit the quantized prediction errors (innovations) to the FC rather than the quantized observations. The analytic and numerical results show that transmitting the innovations rather than the observations mitigates the effect of quantization noise and hence reduces the MSE.