A finite switchboard state machine is a specialized finite state machine. It is built by binding the concepts of switching state machines and commutative state machines. The main purpose of this paper is to give a specific algorithm for fuzzy finite switchboard state machine and also, investigates the concepts of switching relation, covering, restricted cascade products and wreath products of f...

Asynchronous distributed machine learning solutions have proven very effective so far, but always assuming perfectly functioning workers. In practice, some of the workers can however exhibit Byzantine behavior, caused by hardware failures, software bugs, corrupt data, or even malicious attacks. We introduce Kardam, the first distributed asynchronous stochastic gradient descent (SGD) algorithm t...

The problem of counteracting the effects of adversarial inputs on the operation of an asynchronous sequential machine is considered. The objective is to build an automatic state-feedback controller that returns an asynchronous sequential machine to its original state, after the machine has undergone a state transition caused by an adversarial input. It is shown that the existence of such a cont...

A control methodology for asynchronous sequential machines is addressed in this paper. The considered machine consists of a number of input/state asynchronous machines, termed submachines, between which asynchronous switching is conducted by the switching signal. The control objective is to design a corrective controller so that the stable-state behavior of the closed-loop system can mimic a re...

This paper presents a corrective control scheme that automatically counteracts the adverse effect of intermittent faults in the logic of asynchronous sequential machines. When an intermittent fault occurs to the machine, a set of state transitions defined in the machine suspends the normal behavior for a finite interval. The objective is to design a corrective controller that makes the closed-l...

Recent work shows that decentralized parallel stochastic gradient decent (D-PSGD) can outperform its centralized counterpart both theoretically and practically. While asynchronous parallelism is a powerful technology to improve the efficiency of parallelism in distributed machine learning platforms and has been widely used in many popular machine learning softwares and solvers based on centrali...

Stochastic Gradient Descent (SGD) is the standard numerical method used to solve the core optimization problem for the vast majority of machine learning (ML) algorithms. In the context of large scale learning, as utilized by many Big Data applications, efficient parallelization of SGD is in the focus of active research. Recently, we were able to show that the asynchronous communication paradigm...