检索结果-内蒙古大学图书馆

Fast federated learning in the presence of arbitrary device unavailability

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Gu, Xinran Zhang, Jingzhao Huang, Kaixuan Huang, Longbo Department of Industrial Engineering Tsinghua University China EECS Massachusetts Institute of Technology ECE Princeton University United States IIIS Tsinghua University China

Federated learning (FL) coordinates with numerous heterogeneous devices to collaboratively train a shared model while preserving user privacy. Despite its multiple advantages, FL faces new challenges. One challenge arises when devices drop out of the training process beyond the control of the central server. In this case, the convergence of popular FL algorithms such as FedAvg is severely influenced by the straggling devices. To tackle this challenge, we study federated learning algorithms under arbitrary device unavailability and propose an algorithm named Memory-augmented Impatient Federated Averaging (MIFA). Our algorithm efficiently avoids excessive latency induced by inactive devices, and corrects the gradient bias using the memorized latest updates from the devices. We prove that MIFA achieves minimax optimal convergence rates on non-i.i.d. data for both strongly convex and non-convex smooth functions. We also provide an explicit characterization of the improvement over baseline algorithms through a case study, and validate the results by numerical experiments on real-world datasets. Copyright © 2021, The Authors. All rights reserved.

关键词： learning algorithms

Laboratory Demonstration of Decentralized, Physics-Driven learning

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Dillavou, Sam Stern, Menachem Liu, Andrea J. Durian, Douglas J. Department of Physics and Astronomy University of Pennsylvania 209 South 33rd Street PhiladelphiaPA19104 United States

In typical artificial neural networks, neurons adjust according to global calculations of a central processor, but in the brain neurons and synapses self-adjust based on local information. Contrastive learning algorithms have recently been proposed to train physical systems, such as fluidic, mechanical, or electrical networks, to perform machine learning tasks from local evolution rules. However, to date such systems have only been implemented in silico due to the engineering challenge of creating elements that autonomously evolve based on their own response to two sets of global boundary conditions. Here we introduce and implement a physics-driven contrastive learning scheme for a network of variable resistors, using circuitry to locally compare the response of two identical networks subjected to the two different sets of boundary conditions. Using this innovation, our system effectively trains itself, optimizing its resistance values without use of a central processor or external information storage. Once the system is trained for a specified allostery, regression, or classification task, the task is subsequently performed rapidly and automatically by the physical imperative to minimize power dissipation in response to the given voltage inputs. We demonstrate that, unlike typical computers, such learning systems are robust to extreme damage (and thus manufacturing defects) due to their decentralized learning. Our twin-network approach is therefore readily scalable to extremely large or nonlinear networks where its distributed nature will be an enormous advantage;a laboratory network of only 500 edges will already outpace its in silico counterpart. Copyright © 2021, The Authors. All rights reserved.

关键词： learning algorithms

Automated machine learning techniques for data streams

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Imbrea, Alexandru-Ionut University of Twente P.O. Box 217 Enschede7500AE Netherlands

Automated Machine learning (AutoML) techniques benefitted from tremendous research progress recently. These developments and the continuous-growing demand for machine learning experts led to the development of numerous AutoML tools. Industry applications of machine learning on streaming data become more popular due to the increasing adoption of real-time streaming in IoT, microservices architectures, web analytics, and other fields. However, the AutoML tools assume that the entire training dataset is available upfront and that the underlying data distribution does not change over time. These assumptions do not hold in a data-stream-mining setting where an unbounded stream of data cannot be stored and is likely to manifest concept drift. This research surveys the state-of-the-art open-source AutoML tools, applies them to real and synthetic streamed data, and measures how their performance changes over time. For comparative purposes, batch, batch incremental and instance incremental estimators are applied and compared. Moreover, a meta-learning technique for online algorithm selection based on meta-feature extraction is proposed and compared, while model replacement and continual AutoML techniques are discussed. The results show that off-the-shelf AutoML tools can provide satisfactory results but in the presence of concept drift, detection or adaptation techniques have to be applied to maintain the predictive accuracy over time. © 2021, CC BY.

关键词： learning algorithms

Hyperparameter selection for imitation learning

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Hussenot, Léonard Andrychowicz, Marcin Vincent, Damien Dadashi, Robert Raichuk, Anton Stafiniak, Lukasz Girgin, Sertan Marinier, Raphael Momchev, Nikola Ramos, Sabela Orsini, Manu Bachem, Olivier Geist, Matthieu Pietquin, Olivier Google Research Brain Team Univ. de Lille CNRS Inria Scool UMR 9189 CRIStAL

We address the issue of tuning hyperparameters (HPs) for imitation learning algorithms in the context of continuous-control, when the underlying reward function of the demonstrating expert cannot be observed at any time. The vast literature in imitation learning mostly considers this reward function to be available for HP selection, but this is not a realistic setting. Indeed, would this reward function be available, it could then directly be used for policy training and imitation would not be necessary. To tackle this mostly ignored problem, we propose a number of possible proxies to the external reward. We evaluate them in an extensive empirical study (more than 10'000 agents across 9 environments) and make practical recommendations for selecting HPs. Our results show that while imitation learning algorithms are sensitive to HP choices, it is often possible to select good enough HPs through a proxy to the reward function. © 2021, CC BY-NC-SA.

关键词： learning algorithms

IMPROVING FAIRNESS VIA FEDERATED learning

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Zeng, Yuchen Chen, Hongxu Lee, Kangwook University of Wisconsin-Madison United States

Recently, lots of algorithms have been proposed for learning a fair classifier from decentralized data. However, many theoretical and algorithmic questions remain open. First, is federated learning necessary, i.e., can we simply train locally fair classifiers and aggregate them? In this work, we first propose a new theoretical framework, with which we demonstrate that federated learning can strictly boost model fairness compared with such non-federated algorithms. We then theoretically and empirically show that the performance tradeoff of FEDAVG-based fair learning algorithms is strictly worse than that of a fair classifier trained on centralized data. To bridge this gap, we propose FEDFB, a private fair learning algorithm on decentralized data. The key idea is to modify the FEDAVG protocol so that it can effectively mimic the centralized fair learning. Our experimental results show that FEDFB significantly outperforms existing approaches, sometimes matching the performance of the centrally trained model. Copyright © 2021, The Authors. All rights reserved.

关键词： learning algorithms

Tuning the Weights: The Impact of Initial Matrix Configurations on Successor Features learning Efficacy

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Lee, Hyunsu The Faculty of Department of Physiology School of Medicine Pusan National University Yangsan Korea Republic of

The focus of this study is to investigate the impact of different initialization strategies for the weight matrix of Successor Features (SF) on learning efficiency and convergence in Reinforcement learning (RL) agents. Using a grid-world paradigm, we compare the performance of RL agents, whose SF weight matrix is initialized with either an identity matrix, zero matrix, or a randomly generated matrix (using Xavier, He, or uniform distribution method). Our analysis revolves around evaluating metrics such as value error, step length, PCA of Successor Representation (SR) place field, and the distance of SR matrices between different agents. The results demonstrate that RL agents initialized with random matrices reach the optimal SR place field faster and showcase a quicker reduction in value error, pointing to more efficient learning. Furthermore, these random agents also exhibit a faster decrease in step length across larger grid-world environments. The study provides insights into the neurobiological interpretations of these results, their implications for understanding intelligence, and potential future research directions. These findings could have profound implications for the field of artificial intelligence, particularly in the design of learning algorithms. © 2021, CC BY-NC-ND.

关键词： learning algorithms

Stragglers are not disaster: A hybrid federated learning algorithm with delayed gradients

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Li, Xingyu Qu, Zhe Tang, Bo Lu, Zhuo

Federated learning (FL) is a new machine learning framework which trains a joint model across a large amount of decentralized computing devices. Existing methods, e.g., Federated Averaging (FedAvg), are able to provide an optimization guarantee by synchronously training the joint model, but usually suffer from stragglers, i.e., IoT devices with low computing power or communication bandwidth, especially on heterogeneous optimization problems. To mitigate the influence of stragglers, this paper presents a novel FL algorithm, namely Hybrid Federated learning (HFL), to achieve a learning balance in efficiency and effectiveness. It consists of two major components: synchronous kernel and asynchronous updater. Unlike traditional synchronous FL methods, our HFL introduces the asynchronous updater which actively pulls unsynchronized and delayed local weights from stragglers. An adaptive approximation method, Adaptive Delayed-SGD (AD-SGD), is proposed to merge the delayed local updates into the joint model. The theoretical analysis of HFL shows that the convergence rate of the proposed algorithm is O(t+1τ ) for both convex and non-convex optimization problems. Copyright © 2021, The Authors. All rights reserved.

关键词： learning algorithms

Enhanced Membership Inference Attacks against Machine learning Models

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Ye, Jiayuan Maddi, Aadyaa Murakonda, Sasi Kumar Bindschaedler, Vincent Shokri, Reza National University of Singapore Singapore Privitar Labs University of Florida United States

How much does a machine learning algorithm leak about its training data, and why? Membership inference attacks are used as an auditing tool to quantify this leakage. In this paper, we present a comprehensive hypothesis testing framework that enables us not only to formally express the prior work in a consistent way, but also to design new membership inference attacks that use reference models to achieve a significantly higher power (true positive rate) for any (false positive rate) error. More importantly, we explain why different attacks perform differently. We present a template for indistinguishability games, and provide an interpretation of attack success rate across different instances of the game. We discuss various uncertainties of attackers that arise from the formulation of the problem, and show how our approach tries to minimize the attack uncertainty to the one bit secret about the presence or absence of a data point in the training set. We perform a differential analysis between all types of attacks, explain the gap between them, and show what causes data points to be vulnerable to an attack (as the reasons vary due to different granularities of memorization, from overfitting to conditional memorization). Our auditing framework is openly accessible as part of the Privacy Meter software tool. © 2021, CC BY.

关键词： learning algorithms

PowerGym: A Reinforcement learning Environment for Volt-Var Control in Power Distribution Systems

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Fan, Ting-Han Lee, Xian Yeow Wang, Yubo Princeton University United States Iowa State University United States Siemens Technology

We introduce PowerGym, an open-source reinforcement learning environment for Volt-Var control in power distribution systems. Following OpenAI Gym APIs, PowerGym targets minimizing power loss and voltage violations under physical networked constraints. PowerGym provides four distribution systems (13Bus, 34Bus, 123Bus, and 8500Node) based on IEEE benchmark systems and design variants for various control difficulties. To foster generaliza-tion, PowerGym offers a detailed customization guide for users working with their distribution systems. As a demonstration, we examine state-of-the-art reinforcement learning algorithms in PowerGym and validate the environment by studying controller behaviors. The repository is available at https://***/siemens/powergym. Copyright © 2021, The Authors. All rights reserved.

关键词： learning algorithms