19 papers · ranked by Valyu relevance
Mahmoud Assran, Arda Aytekin, Hamid Reza Feyzmahdavian, Mikael Johansson + 1 more
'Mikael Johansson' 'Michael Rabbat'] Motivated by large-scale optimization problems arising in the context of machine learning, there have been several advances in the study of asynchronous parallel and distributed optimization methods during the past decade. Asynchronous methods do not require all processors to…
Saeed Soori, Bugra Can, Mert Gürbüzbalaban, Maryam Mehri Dehnavi
—ASYNC is a framework that supports the implementation of asynchrony and history for optimization methods on distributed computing platforms. The popularity of asynchronous optimization methods has increased in distributed machine learning. However, their applicability and practical experimentation on distributed…
Horia Mania, Xinghao Pan, Dimitris Papailiopoulos, Benjamin Recht + 2 more
'Kannan Ramchandran' 'Michael I. Jordan'] We introduce and analyze stochastic optimization methods where the input to each update is perturbed by bounded noise. We show that this framework forms the basis of a unified approach to analyze asynchronous implementations of stochastic optimization algorithms, by viewing…
Artin Spiridonoff, Alex Olshevsky, Ioannis Ch. Paschalidis
We consider the standard model of distributed optimization of a sum of functions $F(z)=\sum_{i=1}^{n}f_{i}(z)$, where node i in a network holds the function fi(z). We allow for a harsh network model characterized by asynchronous updates, message delays, unpredictable message losses, and directed communication among…
Cong Fang, Yameng Huang, Zhouchen Lin
Asynchronous algorithms have attracted much attention recently due to the crucial demands on solving large-scale optimization problems. However, the accelerated versions of asynchronous algorithms are rarely studied. In this paper, we propose the "momentum compensation" technique to accelerate asynchronous algorithms…
Shay Snyder, Sumedh R. Risbud, Maryam Parsa
Performing optimization with event-based asynchronous neuromorphic systems presents significant challenges. Intel's neuromorphic computing framework, Lava, offers an abstract application programming interface designed for constructing event-based computational graphs. In this study, we introduce a novel framework…
Joseph E. Gonzalez, Peter Bailis, Michael I. Jordan, Michael J. Franklin + 3 more
'Michael J. Franklin' 'Joseph M. Hellerstein' 'Ali Ghodsi' 'Ion Stoica'] Scalable distributed dataflow systems have recently experienced widespread adoption, with commodity dataflow engines such as Hadoop and Spark, and even commodity SQL engines routinely supporting increasingly sophisticated analytics tasks (e.g.…
Riley Hickman, Malcolm Sim, Sergio Pablo-García, Ivan Woolhouse + 6 more
Self-driving laboratories (SDLs) are next-generation research and development platforms for closed-loop, autonomous experimentation that combine ideas from artificial intelligence, robotics, and high-performance computing. A critical component of SDLs is the decision-making algorithm used to prioritize experiments to…
Authors not listed
With the ever-increasing demand for atomistic structures representative of real-life systems as well as the ad-vent of exascale computers, it has now become necessary and possible to use advanced global optimization (GO) techniques to intelligently sample the potential energy surface (PES). Given the previous studies…
Ren Xiaoguang, Xu Xinhai
Computational fluid dynamics (CFD) simulation often needs to periodically output intermediate results to files in the form of snapshots for visualization or restart, which seriously impacts the performance. In this paper, we present asynchronous pipeline I/O (AP-IO) optimization scheme for the periodically snapshot…
Trevor S Frisby, Zhiyun Gong, Christopher James Langmead
Bayesian Optimization is a sequential strategy for optimizing black-box objective functions, f. As mentioned in the introduction, Gaussian Processes (GP) are frequently used to represent and compute over the distribution P(f). A GP is defined by a mean function $μ:X\rightarrowR$ and kernel function…
J Kyle Medley, Shaik Asifullah, Joseph Hellerstein, Herbert M Sauro
Mechanistic kinetic models of biological pathways are an important tool for understanding biological systems. Constructing kinetic models requires fitting the parameters to experimental data. However, parameter fitting on these models is a non–convex, non–linear optimization problem. Many algorithms have been proposed…
Kuiwu Wang, Qin Zhang, Guimei Zheng, Xiaolong Hu + 1 more
Aiming at the problem of asynchronous multi-target tracking, this paper studies the AA fusion optimization problem of multi-sensor networks. Firstly, each sensor node runs a PHD filter, and the measurement information obtained from different sensor nodes in the fusion interval is flood communicated into composite…
Haijie Pan, Lirong Zheng, Sylvain Girard
Machine learning models often converge slowly and are unstable due to the significant variance of random data when using a sample estimate gradient in SGD. To increase the speed of convergence and improve stability, a distributed SGD algorithm based on variance reduction, named DisSAGD, is proposed in this study.…
Zuhan Liu, Kexin Zhao, Xuehu Liu, Huan Xu
With the rapid expansion of industrialization and urbanization, fine Particulate Matter (PM2.5) pollution has escalated into a major global environmental crisis. This pollution severely affects human health and ecosystem stability. Accurately predicting PM2.5 levels is essential. However, air quality forecasting…
Dan Guevarra, Kevin Kan, Yungchieh Lai, Ryan Jones + 5 more
Advancements in artificial intelligence (AI) for science are continually expanding the value proposition for automation in materials and chemistry experiments. The advent of hierarchical decision-making also motivates automation of not only the individual measurements but also the coordination among multiple research…
Sterling Baird, Jason R. Hall, Taylor D. Sparks
Would you rather search for a line inside a cube or a point inside a square? Physics-based simulations and wet-lab experiments often have symmetries (degeneracies) that allow reducing problem dimensionality or search space, but constraining these degeneracies is often unsupported or difficult to implement in many…
Peter L. Bartlett, Chris Junchi Li, Jingfeng Wu, Bin Yu
In the field of optimization, developing accelerated methods for solving minimax and fixed-point problems remains a fundamental challenge. This paper presents a novel family of dual accelerated algorithms that achieve optimal convergence rates for both minimax and fixed-point problems. By exploring new anchoring…
Riley Hickman, Matteo Aldeghi, Alán Aspuru-Guzik
Model-based optimization strategies, such as Bayesian optimization (BO), have been deployed across the natural sciences in design and discovery campaigns due to their sample efficiency and flexibility. The combination of such strategies with automated laboratory equipment and/or high-performance computing in a…