-
Distributed gradient methods under heavy-tailed communication noise
Authors:
Manojlo Vukovic,
Dusan Jakovetic,
Dragana Bajovic,
Soummya Kar
Abstract:
We consider a standard distributed optimization problem in which networked nodes collaboratively minimize the sum of their locally known convex costs. For this setting, we address for the first time the fundamental problem of design and analysis of distributed methods to solve the above problem when inter-node communication is subject to \emph{heavy-tailed} noise. Heavy-tailed noise is highly rele…
▽ More
We consider a standard distributed optimization problem in which networked nodes collaboratively minimize the sum of their locally known convex costs. For this setting, we address for the first time the fundamental problem of design and analysis of distributed methods to solve the above problem when inter-node communication is subject to \emph{heavy-tailed} noise. Heavy-tailed noise is highly relevant and frequently arises in densely deployed wireless sensor and Internet of Things (IoT) networks. Specifically, we design a distributed gradient-type method that features a carefully balanced mixed time-scale time-varying consensus and gradient contribution step sizes and a bounded nonlinear operator on the consensus update to limit the effect of heavy-tailed noise. Assuming heterogeneous strongly convex local costs with mutually different minimizers that are arbitrarily far apart, we show that the proposed method converges to a neighborhood of the network-wide problem solution in the mean squared error (MSE) sense, and we also characterize the corresponding convergence rate. We further show that the asymptotic MSE can be made arbitrarily small through consensus step-size tuning, possibly at the cost of slowing down the transient error decay. Numerical experiments corroborate our findings and demonstrate the resilience of the proposed method to heavy-tailed (and infinite variance) communication noise. They also show that existing distributed methods, designed for finite-communication-noise-variance settings, fail in the presence of infinite variance noise.
△ Less
Submitted 30 May, 2025;
originally announced May 2025.
-
Nonlinear consensus+innovations under correlated heavy-tailed noises: Mean square convergence rate and asymptotics
Authors:
Manojlo Vukovic,
Dusan Jakovetic,
Dragana Bajovic,
Soummya Kar
Abstract:
We consider distributed recursive estimation of consensus+innovations type in the presence of heavy-tailed sensing and communication noises. We allow that the sensing and communication noises are mutually correlated while independent identically distributed (i.i.d.) in time, and that they may both have infinite moments of order higher than one (hence having infinite variances). Such heavy-tailed,…
▽ More
We consider distributed recursive estimation of consensus+innovations type in the presence of heavy-tailed sensing and communication noises. We allow that the sensing and communication noises are mutually correlated while independent identically distributed (i.i.d.) in time, and that they may both have infinite moments of order higher than one (hence having infinite variances). Such heavy-tailed, infinite-variance noises are highly relevant in practice and are shown to occur, e.g., in dense internet of things (IoT) deployments. We develop a consensus+innovations distributed estimator that employs a general nonlinearity in both consensus and innovations steps to combat the noise. We establish the estimator's almost sure convergence, asymptotic normality, and mean squared error (MSE) convergence. Moreover, we establish and explicitly quantify for the estimator a sublinear MSE convergence rate. We then quantify through analytical examples the effects of the nonlinearity choices and the noises correlation on the system performance. Finally, numerical examples corroborate our findings and verify that the proposed method works in the simultaneous heavy-tail communication-sensing noise setting, while existing methods fail under the same noise conditions.
△ Less
Submitted 9 November, 2023; v1 submitted 22 December, 2022;
originally announced December 2022.
-
An overview of Generalised Veltman Semantics
Authors:
Joost J. Joosten,
Jan Mas Rovira,
Luka Mikec,
Mladen Vuković
Abstract:
Interpretability logics are endowed with relational semantics à la Kripke: Veltman semantics. For certain applications though, this semantics is not fine-grained enough. Back in 1992, in the research group of de Jongh, the notion of generalised Veltman semantics emerged to obtain certain non-derivability results as was first presented by Verbrugge ([76]). It has turned out that this semantics has…
▽ More
Interpretability logics are endowed with relational semantics à la Kripke: Veltman semantics. For certain applications though, this semantics is not fine-grained enough. Back in 1992, in the research group of de Jongh, the notion of generalised Veltman semantics emerged to obtain certain non-derivability results as was first presented by Verbrugge ([76]). It has turned out that this semantics has various good properties. In particular, in many cases completeness proofs become simpler and the richer semantics will allow for filtration arguments as opposed to regular Veltman semantics. This paper aims to give an overview of results and applications of Generalised Veltman semantics up to the current date.
△ Less
Submitted 9 July, 2020;
originally announced July 2020.
-
Interpretability logics and generalized Veltman semantics
Authors:
Luka Mikec,
Mladen Vuković
Abstract:
We obtain modal completeness of the interpretability logics ILP_0 and ILR w.r.t. generalized Veltman semantics. Our proofs are based on the notion of smart (full) labels. We also give shorter proofs of completeness w.r.t. generalized semantics for many classical interpretability logics. We obtain decidability and finite model property w.r.t. generalized semantics for ILP_0 and ILR. Finally, we dev…
▽ More
We obtain modal completeness of the interpretability logics ILP_0 and ILR w.r.t. generalized Veltman semantics. Our proofs are based on the notion of smart (full) labels. We also give shorter proofs of completeness w.r.t. generalized semantics for many classical interpretability logics. We obtain decidability and finite model property w.r.t. generalized semantics for ILP_0 and ILR. Finally, we develop a construction that might be useful for proofs of completeness of extensions of ILW w.r.t. generalized semantics in the future, and demonstrate its usage with ILW* = ILM_0W.
△ Less
Submitted 8 July, 2019;
originally announced July 2019.
-
Complexity of the interpretability logic IL
Authors:
Luka Mikec,
Fedor Pakhomov,
Mladen Vuković
Abstract:
We show that the decision problem for the basic system of interpretability logic IL is PSPACE-complete. For this purpose we present an algorithm which uses polynomial space with respect to the complexity of a given formula. The existence of such algorithm, together with the previously known PSPACE hardness of the closed fragment of IL, implies PSPACE-completeness.
We show that the decision problem for the basic system of interpretability logic IL is PSPACE-complete. For this purpose we present an algorithm which uses polynomial space with respect to the complexity of a given formula. The existence of such algorithm, together with the previously known PSPACE hardness of the closed fragment of IL, implies PSPACE-completeness.
△ Less
Submitted 6 April, 2018; v1 submitted 16 October, 2017;
originally announced October 2017.
-
Inversion Theorem for Bilinear Hilbert Transform
Authors:
A. Bučkovska,
S. Pilipović,
M. Vuković
Abstract:
An approximation result for the bilinear Hilbert transform is proved and used for the inversion of the bilinear Hilbert transform. Also, p-Lebesgue points $(p\geq 1)$ are analyzed.
An approximation result for the bilinear Hilbert transform is proved and used for the inversion of the bilinear Hilbert transform. Also, p-Lebesgue points $(p\geq 1)$ are analyzed.
△ Less
Submitted 26 February, 2007;
originally announced February 2007.
