Communication, Control and Signal Processing Seminar
Abstracts of talks
02/19 Alexey Frolov (IITP RAS, Moscow), On Multi-Threshold Decoding of q-ary LDPC codes
ABSTRACT: We consider decoding of LDPC codes over GF(q) with a low-complexity majority algorithm. A modification of this algorithm with multiple thresholds is suggested. A lower estimate on the decoding radius realized by the new algorithm is derived. The estimate is shown to be better than the estimate for a single threshold majority decoder. At the same time, introducing multiple thresholds does not affect the order of decoding complexity.
02/26 Alborz Alavian (UMD), Stabilizing Decentralized Systems with Arbitrary Information Structure
A seminal result in decentralized control is the development of fixed modes by Wang
and Davison in 1973 - that plant modes which cannot be moved with a static decentralized
controller cannot be moved by a dynamic one either, and that the other modes which can be
moved can be shifted to any chosen location with arbitrary precision.
These results were developed for perfectly decentralized, or block diagonal, information structure,
where each control input may only depend on a single corresponding measurement. Furthermore,
the results were claimed after a preliminary step was demonstrated, omitting a rigorous induction
for each of these results, and the remaining task is nontrivial.
In this paper, we consider fixed modes for arbitrary information structures, where certain control inputs may depend on some measurements but not others. We provide a comprehensive proof that the modes which cannot be altered by a static controller with the given structure cannot be moved by a dynamic one either. Furthermore, the modes which can be altered by a static controller with the given structure can be moved by a dynamic one to any chosen location with arbitrary precision, thus generalizing and solidifying the Wang and Davison's result.
This shows that a system can be stabilized by a linear time-invariant controller with the given information structure as long as all of the modes which are fixed with respect to that structure are in the left half-plane; an algorithm for synthesizing such a stabilizing decentralized controller is then distilled from the proof.
03/12 Sikai Qu (UMD), Zero-one law for the absence of isolated nodes for Multiplicative Attributes Graph
ABSTRACT: Graphs are powerful ways to describe social, technological, and biological systems with multiple agents involved, where nodes represent agents (people, websites, gene, etc.) and edges (or links) represent relations (or interactions) between agents. Particularly, in the context of social networks if links establishment is a joint effect of features or characteristics of mutual parties, then links automatically serves as causal factors of various social behaviors, phenomena and outcomes. In a recently introduced social network model, it assigned (random) attributes to every nodes and the randomness of links formation depends on attributes of two end points. It is known as Multiplicative Attributes Graph (MAG). Issues regarding model properties, such as connectivity, emergence of triangles, diameter and degree distributions are of interest to study. In this talk, I will formally address the MAG model and give a proof of the zero-one law for the absence of isolated nodes.
03/26 Janis Notzel (Technische Universitat Munchen), Arbitrarily Varying Wiretap Channel
ABSTRACT: We give an overview over recent developments concerning arbitrarily varying wiretap channels. Especially, we are concerned with the role that common randomness (CR) plays for this model. We start with the case where no CR is available. Here, we observe super-activation of the corresponding capacity as well as discontinuities. We characterize the points of discontinuity. Then, we switch attention to the case where small amounts of CR are available. We state a corresponding capacity formula. In that regime it is of little importance whether the Eavesdropper gets to know the CR or not. If one wants to benefit from large amounts of CR, it has to be kept secret from the Eavesdropper. A capacity formula for this case is given as well. In both cases, the capacity formulas are given by regularized expressions.
04/02 Abbas Kazemipour (UMD), Robust Estimation of Self-Exciting Point Process with Application to Neuronal Modeling
ABSTRACT: Binary data can be used to model spontaneous activities in many real-world applications. Applications include neural spiking activities, seismology, criminology, psyology, finance etc. Modeling, robust estimation and prediction of such data then becomes a problem of interest. It turns out that most of the time, future evolution of real-world data depends highly on their history. In this talk, the performance of l1-regularized maximum likelihood estimator is investigated in sparse estimation of self-exciting point processes with highly dependent samples, including the Hawkes process. As we will show such estimation will overcome the shortcomings of standard maximum likelihood method. Non-asymptotic error bounds and convergence rates will be given as well statistical tests of goodness-of-fit. A significant gap in mean squared error of the regularized and maximum likelihood estimates is observed. Application of the regularized estimator to the LGN neuron spiking data will be given where we successfully recovered their intrinsic oscillatory properties.
04/13 Arya Mazumdar (UMN), Can we have local repair over a network?
ABSTRACT: We introduce a model of a distributed storage system that is locally recoverable from any single server failure. Unlike the usual local recovery model of codes for distributed storage, this model accounts for the fact that each server or storage node in a network is connectible to only some, and not all other, nodes. This may happen for reasons such as physical separation, inhomogeneity in storage platforms etc. We estimate the storage capacity of a network under this model and propose some constructive schemes. Some nice connections to other graph-problems and generalizations are also provided.
04/23 Maice Costa (UMD), On the characterization of the age of information
Age of information is an emerging concept, proposed to characterize the timeliness of
information. The development of metrics to analytically characterize the effect of an outdated
message on the performance of a system will certainly benefit the analysis and optimization of
communication networks and control systems.
In this work, we consider status update systems, in which information about the status of a process of interest is transmitted through a network. At the destination node, an action is taken based on the received information. Hence, the timeliness of the message is an important and often critical objective. This is the case, for example, in sensor networks, where sensor nodes report the observations to a central processor. Another example is the use of a feedback channel to report channel side information in mobile networks.
Our work provides contributions to the initial steps in the characterization of the age of information when the source node can manage the samples of a process of interest, deciding to discard or to transmit them to the destination. We model the network using queuing theory, and analyze the age of information under different policies of packet transmission management.