-
Conformal Vector Fields on Complex Hyperbolic Space
Authors:
Hiroyasu Satoh,
Hemangi Madhusudan Shah
Abstract:
In this article, we show that there does not exist any non-trivial conformal vector field on complex hyperbolic space. That is, we show that any conformal vector field on a complex hyperbolic space is indeed a Killing vector field.
In this article, we show that there does not exist any non-trivial conformal vector field on complex hyperbolic space. That is, we show that any conformal vector field on a complex hyperbolic space is indeed a Killing vector field.
△ Less
Submitted 11 June, 2025;
originally announced June 2025.
-
Some Characteristics of Almost $ω$-Bach Solitons
Authors:
Paritosh Ghosh,
Hemangi Madhusudan Shah,
Arindam Bhattacharyya
Abstract:
In this article, we introduce $ω$-Bach tensor corresponding to one form $ω$ and correspondingly introduce almost $ω$-Bach solitons, thereby generalizing the existing notion of Bach tensor and almost Bach solitons. We characterize almost $ω$-Bach solitons, when the potential vector field of the soliton generates an infinitesimal harmonic transformation or is an affine conformal vector field, or is…
▽ More
In this article, we introduce $ω$-Bach tensor corresponding to one form $ω$ and correspondingly introduce almost $ω$-Bach solitons, thereby generalizing the existing notion of Bach tensor and almost Bach solitons. We characterize almost $ω$-Bach solitons, when the potential vector field of the soliton generates an infinitesimal harmonic transformation or is an affine conformal vector field, or is a projective vector field or is a Killing vector field, when the $ω$-Bach tensor is divergence free, or is a harmonic $1$ form or is a Killing $1$-form. We generalize some of the results obtained by P. T. Ho and A. Ghosh. One of the main results of this paper is that we explicitly find some of the gradient almost $ω$-Bach solitons on the product manifolds ${\mathbb S}^2\times{\mathbb H}^2$, $\mathbb{R}^2\times{\mathbb H}^2$ and $\mathbb{R}^2\times{\mathbb S}^2$. Our gradient almost $ω$-Bach solitons generalize the almost Bach solitons on $\mathbb{R}^2\times{\mathbb H}^2$ and $\mathbb{R}^2\times{\mathbb S}^2$ found by P. T. Ho. Moreover, finding of our gradient almost $ω$-Bach solitons on ${\mathbb S}^2\times{\mathbb H}^2$ is a novel one and complements to the existing almost Bach solitons described by P. T. Ho.
△ Less
Submitted 30 May, 2025;
originally announced May 2025.
-
The Funk-Finsler Structure in the Constant Curvature Spaces
Authors:
Ashok Kumar,
Hemangi Madhusudan Shah,
Bankteshwar Tiwari
Abstract:
In this paper, we {\it find} the infinitesimal structure of Funk-Finsler metric in spaces of constant curvature. We investigate the geometry of this Funk-Finsler metric by explicitly computing its $S$-curvature, Riemann curvature, Ricci curvature, and flag curvature. Moreover, we show that the $S$-curvature of the Funk-Finsler metric in hyperbolic space is bounded above by $\frac{3}{2}$, in spheri…
▽ More
In this paper, we {\it find} the infinitesimal structure of Funk-Finsler metric in spaces of constant curvature. We investigate the geometry of this Funk-Finsler metric by explicitly computing its $S$-curvature, Riemann curvature, Ricci curvature, and flag curvature. Moreover, we show that the $S$-curvature of the Funk-Finsler metric in hyperbolic space is bounded above by $\frac{3}{2}$, in spherical space bounded below by $\frac{3}{2}$, and in Euclidean case it is identically equal to $\frac{3}{2}$. Further, we show that the flag curvature of the Funk-Finsler metric in hyperbolic space is bounded above by $-\frac{1}{4}$, in spherical space bounded below by $-\frac{1}{4}$, and in Euclidean case it is identically equal to $-\frac{1}{4}$.
△ Less
Submitted 22 February, 2025;
originally announced February 2025.
-
A Note on Exact State Visit Probabilities in Two-State Markov Chains
Authors:
Mohammad Taha Shah
Abstract:
In this note we derive the exact probability that a specific state in a two-state Markov chain is visited exactly $k$ times after $N$ transitions. We provide a closed-form solution for $\mathbb{P}(N_l = k \mid N)$, considering initial state probabilities and transition dynamics. The solution corrects and extends prior incomplete results, offering a rigorous framework for enumerating state transiti…
▽ More
In this note we derive the exact probability that a specific state in a two-state Markov chain is visited exactly $k$ times after $N$ transitions. We provide a closed-form solution for $\mathbb{P}(N_l = k \mid N)$, considering initial state probabilities and transition dynamics. The solution corrects and extends prior incomplete results, offering a rigorous framework for enumerating state transitions. Numerical simulations validate the derived expressions, demonstrating their applicability in stochastic modeling.
△ Less
Submitted 6 February, 2025; v1 submitted 5 February, 2025;
originally announced February 2025.
-
The Funk-Finsler structure on the unit disc in the hyperbolic plane
Authors:
Ashok Kumar,
Hemangi Madhusudan Shah,
Bankteshwar Tiwari
Abstract:
In this paper, we construct the Funk-Finsler structure in various models of the hyperbolic plane. In particular, in the unit disc of the Klein model, it turns out to be a Randers metric, which is a non-Berwald Douglas metric. Further, using Finsler isometries we obtain the Funk-Finsler structures in other models of the hyperbolic plane. Finally, we also investigate the geometry of this Funk-Finsle…
▽ More
In this paper, we construct the Funk-Finsler structure in various models of the hyperbolic plane. In particular, in the unit disc of the Klein model, it turns out to be a Randers metric, which is a non-Berwald Douglas metric. Further, using Finsler isometries we obtain the Funk-Finsler structures in other models of the hyperbolic plane. Finally, we also investigate the geometry of this Funk-Finsler metric by explicitly computing the S-curvature, Riemann curvature, flag curvature, and Ricci curvature in the Klein unit disc.
△ Less
Submitted 10 January, 2024;
originally announced January 2024.
-
Adaptive Consensus: A network pruning approach for decentralized optimization
Authors:
Suhail M. Shah,
Albert S. Berahas,
Raghu Bollapragada
Abstract:
We consider network-based decentralized optimization problems, where each node in the network possesses a local function and the objective is to collectively attain a consensus solution that minimizes the sum of all the local functions. A major challenge in decentralized optimization is the reliance on communication which remains a considerable bottleneck in many applications. To address this chal…
▽ More
We consider network-based decentralized optimization problems, where each node in the network possesses a local function and the objective is to collectively attain a consensus solution that minimizes the sum of all the local functions. A major challenge in decentralized optimization is the reliance on communication which remains a considerable bottleneck in many applications. To address this challenge, we propose an adaptive randomized communication-efficient algorithmic framework that reduces the volume of communication by periodically tracking the disagreement error and judiciously selecting the most influential and effective edges at each node for communication. Within this framework, we present two algorithms: Adaptive Consensus (AC) to solve the consensus problem and Adaptive Consensus based Gradient Tracking (AC-GT) to solve smooth strongly convex decentralized optimization problems. We establish strong theoretical convergence guarantees for the proposed algorithms and quantify their performance in terms of various algorithmic parameters under standard assumptions. Finally, numerical experiments showcase the effectiveness of the framework in significantly reducing the information exchange required to achieve a consensus solution.
△ Less
Submitted 5 September, 2023;
originally announced September 2023.
-
Asymptotically harmonic manifolds of dimension 3 with minimal horospheres
Authors:
Jihun Kim,
JeongHyeong Park,
Hemangi Madhusudan Shah
Abstract:
In [14], it was shown that, if M is a 3-dimensional asymptotically harmonic with minimal horospheres, then M is flat. However, there is a gap in the proof of this paper. In this paper, we provide the correct proof of the result. Thus we complete the classification of asymptotically harmonic manifolds of dimension 3: An asymptotically harmonic manifold of dimension 3 is either a flat or real hyperb…
▽ More
In [14], it was shown that, if M is a 3-dimensional asymptotically harmonic with minimal horospheres, then M is flat. However, there is a gap in the proof of this paper. In this paper, we provide the correct proof of the result. Thus we complete the classification of asymptotically harmonic manifolds of dimension 3: An asymptotically harmonic manifold of dimension 3 is either a flat or real hyperbolic space.
△ Less
Submitted 5 September, 2023;
originally announced September 2023.
-
A Stochastic Gradient Tracking Algorithm for Decentralized Optimization With Inexact Communication
Authors:
Suhail M. Shah,
Raghu Bollapragada
Abstract:
Decentralized optimization is typically studied under the assumption of noise-free transmission. However, real-world scenarios often involve the presence of noise due to factors such as additive white Gaussian noise channels or probabilistic quantization of transmitted data. These sources of noise have the potential to degrade the performance of decentralized optimization algorithms if not effecti…
▽ More
Decentralized optimization is typically studied under the assumption of noise-free transmission. However, real-world scenarios often involve the presence of noise due to factors such as additive white Gaussian noise channels or probabilistic quantization of transmitted data. These sources of noise have the potential to degrade the performance of decentralized optimization algorithms if not effectively addressed. In this paper, we focus on the noisy communication setting and propose an algorithm that bridges the performance gap caused by communication noise while also mitigating other challenges like data heterogeneity. We establish theoretical results of the proposed algorithm that quantify the effect of communication noise and gradient noise on the performance of the algorithm. Notably, our algorithm achieves the optimal convergence rate for minimizing strongly convex, smooth functions in the context of inexact communication and stochastic gradients. Finally, we illustrate the superior performance of the proposed algorithm compared to its state-of-the-art counterparts on machine learning problems using MNIST and CIFAR-10 datasets.
△ Less
Submitted 27 July, 2023;
originally announced July 2023.
-
Isometric models of the Funk disc and the Busemann function
Authors:
Ashok Kumar,
Hemangi Madhusudan Shah,
Bankteshwar Tiwari
Abstract:
In this article, we find three isometric models of the Funk disc: Finsler upper half of the hyperboloid of two sheets model, the Finsler band model and the Finsler upper hemi sphere model; and we also find two new models of the Finsler-Poincaré disc. We explicitly describe the geodesics in each model. Moreover, we compute the Busemann function and consequently describe the horocycles in the Funk a…
▽ More
In this article, we find three isometric models of the Funk disc: Finsler upper half of the hyperboloid of two sheets model, the Finsler band model and the Finsler upper hemi sphere model; and we also find two new models of the Finsler-Poincaré disc. We explicitly describe the geodesics in each model. Moreover, we compute the Busemann function and consequently describe the horocycles in the Funk and the Hilbert disc. Finally, we prove the asymptotic harmonicity of the Funk disc. We also show that, the concept of asymptotic harmonicity of the Finsler manifolds {\it tacitly} depends on the measure, in {\it contrast} to the Riemannian case.
△ Less
Submitted 10 June, 2023;
originally announced June 2023.
-
Nature of Some Solitons on Almost coKähler Manifolds and Asymptotically Harmonic Manifolds
Authors:
Paritosh Ghosh,
Hemangi Madhusudan Shah,
Arindam Bhattacharyya
Abstract:
In this research, we study the nature of $η$-Einstein and gradient $η$-Einstein soliton in the framework of almost coKähler manifolds and $(κ, μ)$-almost coKähler manifolds. We find some expressions for scalar curvature of the almost coKähler manifold admitting $η$-Einstein soliton in various cases. We also prove that if a $(κ, μ)$-almost coKähler manifold admits a gradient $η$-Einstein soliton, t…
▽ More
In this research, we study the nature of $η$-Einstein and gradient $η$-Einstein soliton in the framework of almost coKähler manifolds and $(κ, μ)$-almost coKähler manifolds. We find some expressions for scalar curvature of the almost coKähler manifold admitting $η$-Einstein soliton in various cases. We also prove that if a $(κ, μ)$-almost coKähler manifold admits a gradient $η$-Einstein soliton, then either the manifold is coKähler, or $N(κ)$-almost coKahler, or the soliton is trivial. We present an example which validates our results. Finally, we investigate the asymptotically harmonic manifolds admitting non-trivial Ricci solitons and show that they exhibit rigid behavior if for example, scalar curvature attains maximum.
△ Less
Submitted 11 May, 2023;
originally announced May 2023.
-
The Isoperimetric Problem in Randers Poincaré Disc
Authors:
Arti Sahu Gangopadhyay,
Ranadip Gangopadhyay,
Hemangi Madhusudan Shah,
Bankteshwar Tiwari
Abstract:
It is known that a simply connected Riemann surface satisfies the isoperimetric equality if and only if it has constant Gaussian curvature. In this article, we show that Randers Poincaré disc satisfies the isoperimetric equality with respect to different volume forms. These metrics do not necessarily have constant (negative) flag curvature. Thus we show that the Osserman's result \cite{RO} in the…
▽ More
It is known that a simply connected Riemann surface satisfies the isoperimetric equality if and only if it has constant Gaussian curvature. In this article, we show that Randers Poincaré disc satisfies the isoperimetric equality with respect to different volume forms. These metrics do not necessarily have constant (negative) flag curvature. Thus we show that the Osserman's result \cite{RO} in the Riemannian case can not be extended to the corresponding Finslerian case.
△ Less
Submitted 27 March, 2023;
originally announced March 2023.
-
An isoperimetric inequality for the harmonic mean of Steklov eigenvalues in rank one symmetric spaces
Authors:
Hemangi Madhusudan Shah,
Sheela Verma
Abstract:
In this note, an isoperimetric inequality for the harmonic mean of lower order Steklov eigenvalues is proved on bounded domains in noncompact rank-$1$ symmetric spaces. This work extends result of \cite{BR.01} and \cite{V.21} proved for bounded domains in Euclidean space and Hyperbolic space, respectively.
In this note, an isoperimetric inequality for the harmonic mean of lower order Steklov eigenvalues is proved on bounded domains in noncompact rank-$1$ symmetric spaces. This work extends result of \cite{BR.01} and \cite{V.21} proved for bounded domains in Euclidean space and Hyperbolic space, respectively.
△ Less
Submitted 11 February, 2023;
originally announced February 2023.
-
Some Solitons on Homogeneous Almost $α$-Cosymplectic $3$-Manifolds and Harmonic Manifolds
Authors:
Naeem Ahmad Pundeer,
Paritosh Ghosh,
Hemangi Madhusudan Shah,
Arindam Bhattacharyya
Abstract:
In this paper, we investigate the nature of Einstein solitons, whether it is steady, shrinking or expanding on almost $α$-cosymplectic $3$-manifolds. We also prove that a simply connected homogeneous almost $α$-cosymplectic $3$-manifold, admitting a contact Einstein soliton, is an unimodular semidirect product Lie group. Finally, we show that a harmonic manifold admits a Ricci soliton if and only…
▽ More
In this paper, we investigate the nature of Einstein solitons, whether it is steady, shrinking or expanding on almost $α$-cosymplectic $3$-manifolds. We also prove that a simply connected homogeneous almost $α$-cosymplectic $3$-manifold, admitting a contact Einstein soliton, is an unimodular semidirect product Lie group. Finally, we show that a harmonic manifold admits a Ricci soliton if and only if it is flat.
△ Less
Submitted 30 January, 2023; v1 submitted 6 January, 2023;
originally announced January 2023.
-
Efficient Distribution Similarity Identification in Clustered Federated Learning via Principal Angles Between Client Data Subspaces
Authors:
Saeed Vahidian,
Mahdi Morafah,
Weijia Wang,
Vyacheslav Kungurtsev,
Chen Chen,
Mubarak Shah,
Bill Lin
Abstract:
Clustered federated learning (FL) has been shown to produce promising results by grouping clients into clusters. This is especially effective in scenarios where separate groups of clients have significant differences in the distributions of their local data. Existing clustered FL algorithms are essentially trying to group together clients with similar distributions so that clients in the same clus…
▽ More
Clustered federated learning (FL) has been shown to produce promising results by grouping clients into clusters. This is especially effective in scenarios where separate groups of clients have significant differences in the distributions of their local data. Existing clustered FL algorithms are essentially trying to group together clients with similar distributions so that clients in the same cluster can leverage each other's data to better perform federated learning. However, prior clustered FL algorithms attempt to learn these distribution similarities indirectly during training, which can be quite time consuming as many rounds of federated learning may be required until the formation of clusters is stabilized. In this paper, we propose a new approach to federated learning that directly aims to efficiently identify distribution similarities among clients by analyzing the principal angles between the client data subspaces. Each client applies a truncated singular value decomposition (SVD) step on its local data in a single-shot manner to derive a small set of principal vectors, which provides a signature that succinctly captures the main characteristics of the underlying distribution. This small set of principal vectors is provided to the server so that the server can directly identify distribution similarities among the clients to form clusters. This is achieved by comparing the similarities of the principal angles between the client data subspaces spanned by those principal vectors. The approach provides a simple, yet effective clustered FL framework that addresses a broad range of data heterogeneity issues beyond simpler forms of Non-IIDness like label skews. Our clustered FL approach also enables convergence guarantees for non-convex objectives. Our code is available at https://github.com/MMorafah/PACFL.
△ Less
Submitted 21 September, 2022;
originally announced September 2022.
-
Outage Analysis of Energy Efficiency in a Finite-Element-IRS Aided Communication System
Authors:
Aaqib Bulla,
Shahid M Shah
Abstract:
In this paper, we study the performance of an energy efficient wireless communication system, assisted by a finite-element-intelligent reflecting surface (IRS). With no instantaneous channel state information (CSI) at the transmitter, we characterize the system performance in terms of the outage probability (OP) of energy efficiency (EE). Depending upon the availability of line-of-sight (LOS) path…
▽ More
In this paper, we study the performance of an energy efficient wireless communication system, assisted by a finite-element-intelligent reflecting surface (IRS). With no instantaneous channel state information (CSI) at the transmitter, we characterize the system performance in terms of the outage probability (OP) of energy efficiency (EE). Depending upon the availability of line-of-sight (LOS) paths, we analyze the system for two different channel models, viz. Rician and Rayleigh. For an arbitrary number of IRS elements $(N)$, we derive the approximate closed-form solutions for the OP of EE, using Laguerre series and moment matching methods. The analytical results are validated using the Monte-Carlo simulations. Moreover, we also quantify the rate of convergence of the derived expressions to the central limit theorem (CLT) approximations using the \textit{Berry-Esseen} inequality. Further, we prove that the OP of EE is a strict pseudo-convex function of the transmit power and hence, has a unique global minimum. To obtain the optimal transmit power, we solve the OP of EE as a constrained optimization problem. To the best of our knowledge, the OP of EE as a performance metric, has never been previously studied in IRS-assisted wireless communication systems.
△ Less
Submitted 17 May, 2022;
originally announced May 2022.
-
The Isoperimetric Problem In Randers Planes
Authors:
Arti Sahu,
Ranadip Gangopadhyay,
Hemangi Madhusudan Shah,
Bankteshwar Tiwari
Abstract:
In this paper, the isoperimetric problem in Randers planes, $(\mathbb{R}^2,F=α+β)$, which are slight deformation of the Euclidean plane $(\mathbb{R}^2,α)$ by suitable one forms $β$, have been studied. We prove that the circles centred at the origin achieves the local maximum area of the isoperimetric problem with respect to well known volume forms in Finsler geometry.
In this paper, the isoperimetric problem in Randers planes, $(\mathbb{R}^2,F=α+β)$, which are slight deformation of the Euclidean plane $(\mathbb{R}^2,α)$ by suitable one forms $β$, have been studied. We prove that the circles centred at the origin achieves the local maximum area of the isoperimetric problem with respect to well known volume forms in Finsler geometry.
△ Less
Submitted 2 April, 2022;
originally announced April 2022.
-
On Minimal Surfaces of Revolutions Immersed in Deformed Hyperbolic Kropina Space
Authors:
Ashok Kumar,
Ranadip Gangopadhyay,
Bankteshwar Tiwari,
Hemangi Madhusudan Shah
Abstract:
In this paper we consider three dimensional upper half space $\mathbb{H}^3 $ equipped with various Kropina metrics obtained by deformation of hyperbolic metric of $\mathbb{H}^3$ through $1$-forms and obtain a partial differential equation that characterizes minimal surfaces immersed in it. We prove that such minimal surfaces can only be obtained when the hyperbolic metric is deformed along $x^3$ d…
▽ More
In this paper we consider three dimensional upper half space $\mathbb{H}^3 $ equipped with various Kropina metrics obtained by deformation of hyperbolic metric of $\mathbb{H}^3$ through $1$-forms and obtain a partial differential equation that characterizes minimal surfaces immersed in it. We prove that such minimal surfaces can only be obtained when the hyperbolic metric is deformed along $x^3$ direction. Then we classify such minimal surfaces and show that flag curvature of these surfaces is always non-positive. We also obtain the geodesics of this surface. In particular, it follows that such surfaces neither have forward conjugate points nor they are forward complete.
△ Less
Submitted 28 February, 2022;
originally announced March 2022.
-
Making Simulated Annealing Sample Efficient for Discrete Stochastic Optimization
Authors:
Suhail M. Shah
Abstract:
We study the regret of simulated annealing (SA) based approaches to solving discrete stochastic optimization problems. The main theoretical conclusion is that the regret of the simulated annealing algorithm, with either noisy or noiseless observations, depends primarily upon the rate of the convergence of the associated Gibbs measure to the optimal states. In contrast to previous works, we show th…
▽ More
We study the regret of simulated annealing (SA) based approaches to solving discrete stochastic optimization problems. The main theoretical conclusion is that the regret of the simulated annealing algorithm, with either noisy or noiseless observations, depends primarily upon the rate of the convergence of the associated Gibbs measure to the optimal states. In contrast to previous works, we show that SA does not need an increased estimation effort (number of \textit{pulls/samples} of the selected \textit{arm/solution} per round for a finite horizon $n$) with noisy observations to converge in probability. By simple modifications, we can make the total number of samples per iteration required for convergence (in probability) to scale as $\mathcal{O}\big(n)$. Additionally, we show that a simulated annealing inspired heuristic can solve the problem of stochastic multi-armed bandits (MAB), by which we mean that it suffers a $\mathcal{O}(\log \,n)$ regret. Thus, our contention is that SA should be considered as a viable candidate for inclusion into the family of efficient exploration heuristics for bandit and discrete stochastic optimization problems.
△ Less
Submitted 26 March, 2021; v1 submitted 14 September, 2020;
originally announced September 2020.
-
Dynamic social learning under graph constraints
Authors:
Konstantin Avrachenkov,
Vivek S. Borkar,
Sharayu Moharir,
Suhail M. Shah
Abstract:
We introduce a model of graph-constrained dynamic choice with reinforcement modeled by positively $α$-homogeneous rewards. We show that its empirical process, which can be written as a stochastic approximation recursion with Markov noise, has the same probability law as a certain vertex reinforced random walk. We use this equivalence to show that for $α> 0$, the asymptotic outcome concentrates aro…
▽ More
We introduce a model of graph-constrained dynamic choice with reinforcement modeled by positively $α$-homogeneous rewards. We show that its empirical process, which can be written as a stochastic approximation recursion with Markov noise, has the same probability law as a certain vertex reinforced random walk. We use this equivalence to show that for $α> 0$, the asymptotic outcome concentrates around the optimum in a certain limiting sense when `annealed' by letting $α\uparrow\infty$ slowly.
△ Less
Submitted 24 July, 2021; v1 submitted 8 July, 2020;
originally announced July 2020.
-
Variable Metric Proximal Gradient Method with Diagonal Barzilai-Borwein Stepsize
Authors:
Youngsuk Park,
Sauptik Dhar,
Stephen Boyd,
Mohak Shah
Abstract:
Variable metric proximal gradient (VM-PG) is a widely used class of convex optimization method. Lately, there has been a lot of research on the theoretical guarantees of VM-PG with different metric selections. However, most such metric selections are dependent on (an expensive) Hessian, or limited to scalar stepsizes like the Barzilai-Borwein (BB) stepsize with lots of safeguarding. Instead, in th…
▽ More
Variable metric proximal gradient (VM-PG) is a widely used class of convex optimization method. Lately, there has been a lot of research on the theoretical guarantees of VM-PG with different metric selections. However, most such metric selections are dependent on (an expensive) Hessian, or limited to scalar stepsizes like the Barzilai-Borwein (BB) stepsize with lots of safeguarding. Instead, in this paper we propose an adaptive metric selection strategy called the diagonal Barzilai-Borwein (BB) stepsize. The proposed diagonal selection better captures the local geometry of the problem while keeping per-step computation cost similar to the scalar BB stepsize i.e. $O(n)$. Under this metric selection for VM-PG, the theoretical convergence is analyzed. Our empirical studies illustrate the improved convergence results under the proposed diagonal BB stepsize, specifically for ill-conditioned machine learning problems for both synthetic and real-world datasets.
△ Less
Submitted 15 October, 2019;
originally announced October 2019.
-
Linearly Convergent Asynchronous Distributed ADMM via Markov Sampling
Authors:
Suhail M. Shah,
Konstantin E. Avrachenkov
Abstract:
We consider the consensual distributed optimization problem and propose an asynchronous version of the Alternating Direction Method of Multipliers (ADMM) algorithm to solve it. The `asynchronous' part here refers to the fact that only one node/processor is updated (i.e. performs a minimization step) at each iteration of the algorithm. The selection of the node to be updated is decided by simulatin…
▽ More
We consider the consensual distributed optimization problem and propose an asynchronous version of the Alternating Direction Method of Multipliers (ADMM) algorithm to solve it. The `asynchronous' part here refers to the fact that only one node/processor is updated (i.e. performs a minimization step) at each iteration of the algorithm. The selection of the node to be updated is decided by simulating a Markov chain. The proposed algorithm is shown to have a linear convergence property in expectation for the class of functions which are strongly convex and continuously differentiable.
△ Less
Submitted 31 March, 2022; v1 submitted 11 October, 2018;
originally announced October 2018.
-
Frank-Wolfe variants for minimization of a sum of functions
Authors:
Suhail M Shah
Abstract:
We propose several variants of the Frank-Wolfe algorithm to minimize a sum of functions. The main proposed algorithm is inspired from the dual averaging scheme of Nesterov adapted for Frank Wolfe in a stochastic setting. A distributed version of this scheme is also suggested. Additionally, we propose a Frank-Wolfe variant based on incremental gradient techniques. The convergence rates for all the…
▽ More
We propose several variants of the Frank-Wolfe algorithm to minimize a sum of functions. The main proposed algorithm is inspired from the dual averaging scheme of Nesterov adapted for Frank Wolfe in a stochastic setting. A distributed version of this scheme is also suggested. Additionally, we propose a Frank-Wolfe variant based on incremental gradient techniques. The convergence rates for all the proposed algorithms are established. The performance is studied on least squares regression and multinomial classification.
△ Less
Submitted 10 January, 2023; v1 submitted 25 May, 2018;
originally announced May 2018.
-
Distributed Optimization on Riemannian Manifolds for multi-agent networks
Authors:
Suhail M. Shah
Abstract:
We consider the consensual distributed optimization problem in the Riemannian context. Specifically, the minimization of a sum of functions form is studied where each individual function in the sum is located at the node of a network. An algorithm, which is a direct generalization of the Euclidean case, to solve the problem is proposed. The convergence analysis is carried out in full detail for ge…
▽ More
We consider the consensual distributed optimization problem in the Riemannian context. Specifically, the minimization of a sum of functions form is studied where each individual function in the sum is located at the node of a network. An algorithm, which is a direct generalization of the Euclidean case, to solve the problem is proposed. The convergence analysis is carried out in full detail for geodesically convex as well as non-convex functions. The algorithm is demonstrated using some standard applications which fit the presented framework.
△ Less
Submitted 7 September, 2020; v1 submitted 29 November, 2017;
originally announced November 2017.
-
Stochastic Approximation on Riemannian manifolds
Authors:
Suhail M. Shah
Abstract:
The standard theory of stochastic approximation (SA) is extended to the case when the constraint set is a Riemannian manifold. Specifically, the standard ODE method for analyzing SA schemes is extended to iterations constrained to stay on a manifold using a retraction mapping. In addition, for submanifolds of a Euclidean space, a framework is developed for a projected SA scheme with approximate re…
▽ More
The standard theory of stochastic approximation (SA) is extended to the case when the constraint set is a Riemannian manifold. Specifically, the standard ODE method for analyzing SA schemes is extended to iterations constrained to stay on a manifold using a retraction mapping. In addition, for submanifolds of a Euclidean space, a framework is developed for a projected SA scheme with approximate retractions. The framework is also extended to non-differentiable constraint sets.
△ Less
Submitted 29 November, 2017;
originally announced November 2017.
-
Weighted Singular Value Thresholding and its Application to Background Estimation
Authors:
Aritra Dutta,
Boqing Gong,
Xin Li,
Mubarak Shah
Abstract:
Singular value thresholding (SVT) plays an important role in the well-known robust principal component analysis (RPCA) algorithms which have many applications in computer vision and recommendation systems. In this paper, we formulate and study a weighted singular value thresholding (WSVT) problem, which uses a combination of the nuclear norm and a weighted Frobenius norm. We present an algorithm t…
▽ More
Singular value thresholding (SVT) plays an important role in the well-known robust principal component analysis (RPCA) algorithms which have many applications in computer vision and recommendation systems. In this paper, we formulate and study a weighted singular value thresholding (WSVT) problem, which uses a combination of the nuclear norm and a weighted Frobenius norm. We present an algorithm to numerically solve WSVT and establish the convergence of the algorithm. As a proof of concept, we apply WSVT with a simple choice of weight learned from the data to the background estimation problem in computer vision. The numerical experiments show that our method can outperform RPCA algorithms. This indicates that instead of tackling the computationally expensive $\ell_1$ norm employed in RPCA, one may switch to the weighted Frobenius norm and achieve about the same or even better performance.
△ Less
Submitted 1 July, 2017;
originally announced July 2017.
-
On cyclic associative Abel-Grassman groupoids
Authors:
Muhammad Iqbal,
Imtiaz Ahmad,
Muhammad Shah,
Muhammad Irfan Ali
Abstract:
A new subclass of AG-groupoids, so called, cyclic associative Abel-Grassman groupoids or CA-AG-groupoid is studied. These have been enumerated up to order $6$. A test for the verification of cyclic associativity for an arbitrary AG-groupoid has been introduced. Various properties of CA-AG-groupoids have been studied. Relationship among CA-AG-groupoids and other subclasses of AG-groupoids is invest…
▽ More
A new subclass of AG-groupoids, so called, cyclic associative Abel-Grassman groupoids or CA-AG-groupoid is studied. These have been enumerated up to order $6$. A test for the verification of cyclic associativity for an arbitrary AG-groupoid has been introduced. Various properties of CA-AG-groupoids have been studied. Relationship among CA-AG-groupoids and other subclasses of AG-groupoids is investigated. It is shown that the subclass of CA-AG-groupoid is different from that of the AG{*}-groupoid as well as AG{*}{*}-groupoids.
△ Less
Submitted 5 October, 2015;
originally announced October 2015.
-
Some General Properties of Stein-AG-groupoids and Stein-AG-Test
Authors:
Muhammad Rashad,
Imtiaz Ahmad,
Muhammad Shah,
Amanullah
Abstract:
A groupoid that satisfying the left invertive law is called an AG-groupoid.this concept is extended to introduce a Stein AG-groupoid. We provethe existence by providing some non-associative examples. We also explore some basic and general properties of these AG-groupoids and find their relations with other subclasses of AG-groupoids.
A groupoid that satisfying the left invertive law is called an AG-groupoid.this concept is extended to introduce a Stein AG-groupoid. We provethe existence by providing some non-associative examples. We also explore some basic and general properties of these AG-groupoids and find their relations with other subclasses of AG-groupoids.
△ Less
Submitted 24 June, 2016; v1 submitted 21 March, 2014;
originally announced March 2014.
-
Enumeration of Bi-commutative AG-groupoids
Authors:
Muhammad Rashad,
Imtiaz Ahmad,
Muhammad Shah,
A. B. Saeid
Abstract:
A groupoid satisfying the left invertive law: $ab\cdot c=cb\cdot a$ is called an AG-groupoid and is a generalization of commutative semigroups. We consider the concept of bi-commutativity in AG-groupoids and thus introduce left commutative AG-groupoids, right commutative AG-groupoids and bi-commutative AG-groupoids.
A groupoid satisfying the left invertive law: $ab\cdot c=cb\cdot a$ is called an AG-groupoid and is a generalization of commutative semigroups. We consider the concept of bi-commutativity in AG-groupoids and thus introduce left commutative AG-groupoids, right commutative AG-groupoids and bi-commutative AG-groupoids.
△ Less
Submitted 17 October, 2019; v1 submitted 21 March, 2014;
originally announced March 2014.
-
Some General Properties of LAD and RAD AG-groupoids
Authors:
Muhammad Rashad,
Imtiaz Ahmad,
Muhammad Shah
Abstract:
A groupoid that satisfies the left invertive law: $ab\cdot c=cb\cdot a$ is called an AG-groupoid. We extend the concept of left abelian distributive groupoid (LAD) and right abelian distributive groupoid (RAD) to introduce new subclasses of AG-groupoid, left abelian distributive AG-groupoid and right abelian distributive AG-groupoid. We give their enumeration up to order 6 and find some basic rela…
▽ More
A groupoid that satisfies the left invertive law: $ab\cdot c=cb\cdot a$ is called an AG-groupoid. We extend the concept of left abelian distributive groupoid (LAD) and right abelian distributive groupoid (RAD) to introduce new subclasses of AG-groupoid, left abelian distributive AG-groupoid and right abelian distributive AG-groupoid. We give their enumeration up to order 6 and find some basic relations of these new classes with other known subclasses of AG-groupoids and other relevant algebraic structures. We establish a method to test an arbitrary AG-groupoid for these classes.
△ Less
Submitted 20 March, 2014;
originally announced March 2014.
-
Fuzzy Cosets and Quotient Fuzzy AG-subgroups
Authors:
Amanullah,
Imtiaz Ahmad,
Muhammad Shah
Abstract:
In this paper we extend the concept of fuzzy AG-subgroups. We introduce some results in normal fuzzy AG-subgroups. We define fuzzy cosets and quotient fuzzy AG-subgroups, and prove that the sets of their collection form an AG-subgroup and fuzzy AG-subgroup respectively. We also introduce the fuzzy Lagrange's Theorem of AG-subgroup. It is known that the condition $μ(xy)=μ(yx)$ holds for all $x,y$ i…
▽ More
In this paper we extend the concept of fuzzy AG-subgroups. We introduce some results in normal fuzzy AG-subgroups. We define fuzzy cosets and quotient fuzzy AG-subgroups, and prove that the sets of their collection form an AG-subgroup and fuzzy AG-subgroup respectively. We also introduce the fuzzy Lagrange's Theorem of AG-subgroup. It is known that the condition $μ(xy)=μ(yx)$ holds for all $x,y$ in fuzzy subgroups if $μ$ is normal, but in fuzzy AG-subgroup we show that it holds without normality.
△ Less
Submitted 15 March, 2014;
originally announced March 2014.
-
On Modulo AG-groupoids
Authors:
Amanullah,
Muhammad Rashad,
Imtiaz Ahmad,
Muhammad Shah
Abstract:
A groupoid G is called an AG-groupoid if it satisfies the left invertive law: (ab)c = (cb)a. An AG-group G, is an AG-groupoid with left identity e \in G (that is, ea = a for all a \in G) and for all a \in G there exists a' \in G such that a.a' = a'.a = e. In this article we introduce the concept of AG-groupoids (mod n) and AG-group (mod n) using Vasantha's constructions [1]. This enables us to pro…
▽ More
A groupoid G is called an AG-groupoid if it satisfies the left invertive law: (ab)c = (cb)a. An AG-group G, is an AG-groupoid with left identity e \in G (that is, ea = a for all a \in G) and for all a \in G there exists a' \in G such that a.a' = a'.a = e. In this article we introduce the concept of AG-groupoids (mod n) and AG-group (mod n) using Vasantha's constructions [1]. This enables us to prove that AG-groupoids (mod n) and AG-groups (mod n) exist for every integer n \geq 3. We also give some nice characterizations of some classes of AG-groupoids in terms of AG-groupoids (mod n).
△ Less
Submitted 11 March, 2014;
originally announced March 2014.
-
The Multiplication Group of an AG-group
Authors:
Muhammad Shah,
Asif Ali,
Imtiaz Ahmad,
Volker Sorge
Abstract:
We investigate the multiplication group of a special class of quasigroup called AG-group. We prove some interesting results such as: the multiplication group of an AG-group of order n is non-abelian group of order 2n and its left section is an abelian group of order n. The inner mapping group of an AG-group of any order is a cyclic group of order 2.
We investigate the multiplication group of a special class of quasigroup called AG-group. We prove some interesting results such as: the multiplication group of an AG-group of order n is non-abelian group of order 2n and its left section is an abelian group of order n. The inner mapping group of an AG-group of any order is a cyclic group of order 2.
△ Less
Submitted 3 March, 2014;
originally announced March 2014.
-
Left Transitive AG-groupoids
Authors:
Muhammad Rashad,
Imtiaz Ahmad,
Muhammad Shah,
Z. U. A. Khuhro
Abstract:
An AG-groupoid is an algebraic structure that satisfies the left invertive law: (ab)c =(cb)a. We prove that the class of left transitive AG-groupoids (AG-groupoids satisfying the identity, ab.ac = bc) coincides with the class of T2-AG-groupoids. We also develop a simple procedure to test whether an arbitrary groupoid is left transitive AG-groupoid or not. Further we prove that, (i). Every left tra…
▽ More
An AG-groupoid is an algebraic structure that satisfies the left invertive law: (ab)c =(cb)a. We prove that the class of left transitive AG-groupoids (AG-groupoids satisfying the identity, ab.ac = bc) coincides with the class of T2-AG-groupoids. We also develop a simple procedure to test whether an arbitrary groupoid is left transitive AG-groupoid or not. Further we prove that, (i). Every left transitive AG-groupoid is transitively commutative AG-groupoid (ii) For left transitive AG-groupoid the properties of flexibility, right alternativity, AG*, right nuclear square, middle nuclear square and commutative semigroup are equivalent.
△ Less
Submitted 15 November, 2014; v1 submitted 21 February, 2014;
originally announced February 2014.
-
AG-groups and other classes of right Bol quasigroups
Authors:
Muhammad Shah,
Sergey Shpectorov,
Asif Ali
Abstract:
By a result of Sharma, right Bol quasigroups are obtainable from right Bol loops via an involutive automorphism. We prove that the class of AG-groups, introduced by Kamran, is obtained via the same construction from abelian groups. We further introduce a new class of Bol* quasigroups, which turns out to correspond, as above, to the class of groups.
Sharma's correspondence allows an efficient imp…
▽ More
By a result of Sharma, right Bol quasigroups are obtainable from right Bol loops via an involutive automorphism. We prove that the class of AG-groups, introduced by Kamran, is obtained via the same construction from abelian groups. We further introduce a new class of Bol* quasigroups, which turns out to correspond, as above, to the class of groups.
Sharma's correspondence allows an efficient implementation and we present some enumeration results for the above three classes.
△ Less
Submitted 21 June, 2011; v1 submitted 15 June, 2011;
originally announced June 2011.
