Stochastic gradient descent (SGD) and its variants have become more and more popular in machine learning due to their efficiency and effectiveness. To handle large-scale problems, researchers have recently proposed several parallel SGD methods for multicore systems. However, existing parallel SGD methods cannot achieve satisfactory performance in real applications. In this paper, we propose a f...

The progress of semiconductor technology and control systems the last years has made possible the realisation of highly efficient and low priced microprocessor-based controllers for the control of asynchronous motors. This paper describes the design and realization of a microcomputer-based device for frequency regulation of an asynchronous motor. Through the regulation of motor’s supply frequen...

Asynchrony being consubstantial with the Web, asynchronous communication and control patterns are ubiquitous at multiple scales in Web programs. They are used to program Web clients as well as Web servers. Browser HTML GUIs raise asynchronous events in reaction to user interactions. The Web browser DOM (Document Object Model) creates Web page elements asynchronously. Browser-toserver (AJAX) and...

Large scale multi-agent systems (MASs) in unpredictable environments must use machine learning techniques to perform their goals and improve the performance of the system. This paper presents a systematic approach to introduce machine learning in the design and implementation phases of a software agent. We also present an incremental implementation process for building asynchronous and distribu...

We motivate and present a simple measure of eeciency for distributed data structures called \bottleneck complexity". We investigate a distributed counter as a case study for our measure of eeciency. A distributed counter allows each processor in an asynchronous message passing network to access the counter value and increment it. The bottleneck complexity has a lower bound of (k) on an asynchro...

Distributed computing using networked workstations o ers cost-e cient parallel computing, but the higher rate of failure requires e ective fault-tolerance. Asynchronous consistent checkpointing o ers a low-overhead solution. Parallel Virtual Machine (PVM) allows a heterogeneous network of UNIX workstations to serve immmediately as a distributed computer by providing message-passing services imp...

We introduce and analyze stochastic optimization methods where the input to each gradient updateis perturbed by bounded noise. We show that this framework forms the basis of a unified approachto analyze asynchronous implementations of stochastic optimization algorithms. In this framework,asynchronous stochastic optimization algorithms can be thought of as serial methods operatin...

Zeroth-order (derivative-free) optimization attracts a lot of attention in machine learning, because explicit gradient calculations may be computationally expensive or infeasible. To handle large scale problems both in volume and dimension, recently asynchronous doubly stochastic zeroth-order algorithms were proposed. The convergence rate of existing asynchronous doubly stochastic zeroth order ...

Solvability of the model matching problem for input/output switched asynchronous sequential machines is discussed in this paper. The control objective is to determine the existence condition and design algorithm for a corrective controller that can match the stable-state behavior of the closed-loop system to that of a reference model. Switching operations and correction procedures are incorpora...

In the era of big data, optimizing large scale machine learning problems becomes a challenging task and draws significant attention. Asynchronous optimization algorithms come out as a promising solution. Recently, decoupled asynchronous proximal stochastic gradient descent (DAP-SGD) is proposed to minimize a composite function. It is claimed to be able to offload the computation bottleneck from...