检索结果-内蒙古大学图书馆

GFlowNet Training by Policy Gradients

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Niu, Puhua Wu, Shili Fan, Mingzhou Qian, Xiaoning Department of Electrical and Computer Engineering Texas A&M University College StationTX United States Computational Science Intiative Brookhaven National Laboratory UptonNY United States

Generative Flow Networks (GFlowNets) have been shown effective to generate combinatorial objects with desired properties. We here propose a new GFlowNet training framework, with policy-dependent rewards, that bridges keeping flow balance of GFlowNets to optimizing the expected accumulated reward in traditional Reinforcement-Learning (RL). This enables the derivation of new policy-based GFlowNet training methods, in contrast to existing ones resembling value-based RL. It is known that the design of backward policies in GFlowNet training affects efficiency. We further develop a coupled training strategy that jointly solves GFlowNet forward policy training and backward policy design. Performance analysis is provided with a theoretical guarantee of our policy-based GFlowNet training. Experiments on both simulated and real-world datasets verify that our policy-based strategies provide advanced RL perspectives for robust gradient estimation to improve GFlowNet performance. Our code is available at: ***/niupuhua1234/GFN-PG. © 2024, CC BY.

关键词： Generative adversarial networks

A Case Study on Modeling Adequacy of a Grid with Subsynchronous Oscillations Involving IBRs

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Karunaratne, Lilan Chaudhuri, Nilanjan Ray Yogarathnam, Amirthagunaraj Yue, Meng School of Electrical Engineering and Computer Science Pennsylvania State University University ParkPA United States Interdisciplinary Science Department Brookhaven National Laboratory UptonNY United States

A case study on modeling adequacy of a grid in presence of renewable resources based on grid-forming converters (GFCs) is the subject matter of this paper. For this purpose, a 4-machine 11-bus IEEE benchmark model is modified by considering GFCs replacing synchronous generators that led to unstable subsynchronous oscillations (SSOs). We aim to: (a) understand if transmission network dynamics should be considered in such cases, (b) revisit the space-phasor-calculus (SPC) in d-q frame under balanced condition that captures such phenomena and lends itself to eigenvalue analysis, and (c) emphasize limitations of such models while underscoring their importance for large-scale power system simulations. Time-domain and frequency-domain results from SPC and quasistationary phasor calculus (QPC) models are compared with electromagnetic transient (EMT)based simulations. It is shown that models with transmission line dynamics in SPC framework can capture the SSO mode while QPC models that neglect these dynamics fail to do so. © 2024, CC BY.

关键词： Electric power transmission networks

A Compact Power Stage Design for 10-kV SiC Device-Based Converters with Heatsink Electric Field Shaping

学校读者我要写书评

暂无评论

A Compact Power Stage Design for 10-kV SiC Device-Based Conv...

Annual IEEE Conference on Applied Power Electronics Conference and Exposition (APEC)

作者： Zihan Gao Ruirui Chen Hua Bai Leon Tolbert Fred Wang Min H. Kao Department of Electrical Engineering and Computer Science University of Tennessee Knoxville Knoxville TN USA Oak Ridge National Laboratory Oak Ridge TN USA

The 10-kV SiC device imposes high insulation requirements on the corresponding power stage design. As the voltage stress and clearance distance increases, achieving a compact design of the converter will be challenging. In this paper, a compact power stage design is realized by adopting a new heatsink configuration. With the new heatsink configuration, a field shaping structure is proposed to suppress concentrated electric field close to the edges, so that the clearance between heatsinks can be reduced. Simulation and experimental tests have validated the design. The proposed design has been tested up to 6250 V/10 A, and the gate drive and bus-bar design are also discussed in this paper.

关键词：

Cooperative Multi-Agent Deep Reinforcement Learning for Adaptive Decentralized Emergency Voltage Control

学校读者我要写书评

暂无评论

Cooperative Multi-Agent Deep Reinforcement Learning for Adap...

Innovative Smart Grid Technologies (ISGT)

作者： Ying Zhang Meng Yue Department of Electrical and Computer Engineering Montana State University Bozeman MT U.S. Interdisciplinary Science Department Brookhaven National Laboratory Upton NY U.S.

Under voltage load shedding (UVLS) for power grid emergency control builds the last defensive perimeter to prevent cascade outages and blackouts in case of contingencies. This letter proposes a novel cooperative multi-agent deep reinforcement learning (MADRL)-based UVLS algorithm in an adaptive decentralized way. With well-designed input signals reflecting the voltage deviation, newly structured neural networks are developed as intelligent agents to obtain control actions and their probabilities to accommodate high uncertainties in volatile power system operations. Moreover, the interaction among the agents for coordinated control is implemented and refined by a state-of-the-art attention mechanism, which helps agents concentratively learn effective interacted information. The proposed method realizes decentralized coordinated control, adapting to extremely high uncertainties. Case studies on an IEEE benchmark system indicate the superior performance of the proposed algorithm.

关键词：

6th Place Solution to Google Universal Image Embedding

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Gkelios, Socratis Kastellos, Anestis Chatzichristofis, Savvas A. Department of Electrical and Computer Engineering Democritus University of Thrace Greece Intelligent Systems Laboratory Department of Computer Science Neapolis University Pafos Cyprus

This paper presents the 6th place solution to the Google Universal Image Embedding competition on Kaggle. Our approach is based on the CLIP architecture;a powerful pre-trained model used to learn visual representations from natural language supervision. We also utilized the SubCenter ArcFace loss with dynamic margins to improve the distinctive power of class separability and embeddings. Finally, a diverse dataset has been created based on the test’s set categories and the leaderboard’s feedback. By carefully crafting a training scheme to enhance transfer learning, our submission scored 0.685 on the private leaderboard. Copyright © 2022, The Authors. All rights reserved.

关键词： Embeddings

Matrix Manifold Precoder Design for Massive MIMO High-Speed Railway Communications with Channel Prediction

学校读者我要写书评

暂无评论

Matrix Manifold Precoder Design for Massive MIMO High-Speed ...

International Conference on Communication Technology (ICCT)

作者： Rui Sun Guan-Lin Liu Chen Sun Ding Shi An-An Lu Xiqi Gao Xiang-Gen Xia National Mobile Communications Research Laboratory Southeast University Nanjing China Purple Mountain Laboratories Nanjing China Department of Electrical and Computer Engineering University of Delaware Newark DE USA

ISBN: (数字)9798350363760

ISBN: (纸本)9798350363777

In high-speed railway (HSR) communications, the channel suffers from severe channel aging effect caused by the high mobility. To address this issue, we investigate the precoder design against channel aging in massive multiple-input multiple-output (MIMO) systems with channel prediction. First of all, we introduce the concept of the quadruple beams (QBs), and establish a QB based channel model with sampled quadruple steering vectors. Then, the upcoming space domain channel can achieve a higher accuracy by channel prediction. We consider the precoder design on the Riemannian submanifold formed by the precoders satisfying the total power constraint (TPC). The Riemannian conjugate gradient (RCG) method is proposed to solve the problem on the manifold. The RCG method mainly involves the matrix multiplication and avoids the need of matrix inversion of the transmit antenna dimension. The simulation results demonstrate the effectiveness of the proposed channel model and the superiority of the RCG method for precoder design against channel aging.

关键词： Manifolds Accuracy Simulation Design methodology Transmitting antennas Massive MIMO Aging Vectors Channel models Optimization

Analysis and Applications of a Heralded Electron Source

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Koppell, Stewart A. Simonaitis, John W. Krielaart, Maurice A.R. Berggren, Karl K. Keathley, Phillip D. Putnam, William P. Research Laboratory of Electronics Massachusetts Institute of Technology CambridgeMA02139 United States Department of Electrical and Computer Engineering University of California Davis DavisCA95616 United States

We analytically describe the noise properties of a heralded electron source made from a standard electron gun, a weak photonic coupler, a single photon counter, and an electron energy filter. We argue the traditional heralding figure of merit, the Klyshko efficiency, is an insufficient statistic for characterizing performance in dose-control and dose-limited applications. Instead, we describe the sub-Poissonian statistics of the source using the fractional reduction in variance and the fractional increase in Fisher Information. Using these figures of merit, we discuss the engineering requirements for efficient heralding and evaluate potential applications using simple models of electron lithography, bright-field scanning transmission electron microscopy (BFSTEM), and scanning electron microscopy (SEM). We find that the advantage in each of these applications is situational, but potentially significant: dynamic control of the trade-off between write speed and shot noise in electron lithography;an order of magnitude dose reduction in BFSTEM for thin samples (e.g. 2D materials);and a doubling of dose efficiency for wall-steepness estimation in SEM. © 2024, CC BY.

关键词： Electrons

Residential HVAC Aggregation Based on Risk-averse Multi-armed Bandit Learning for Secondary Frequency Regulation

学校读者我要写书评

暂无评论

Journal of Modern Power Systems and Clean Energy 2020年第6期8卷 1160-1167页

作者： Xinyi Chen Qinran Hu Qingxin Shi Xiangjun Quan Zaijun Wu Fangxing Li the School of Electrical Engineering Southeast UniversityNanjingChina Jiangsu Provincial Key Laboratory of Smart Grid Technology and Equipment NanjingChina the School of Electrical Engineering and Computer Science University of TennesseeKnoxvilleUSA

As the penetration of renewable energy continues to increase,stochastic and intermittent generation resources gradually replace the conventional generators,bringing significant challenges in stabilizing power system ***,aggregating demand-side resources for frequency regulation attracts attentions from both academia and ***,in practice,conventional aggregation approaches suffer from random and uncertain behaviors of the users such as opting out control *** risk-averse multi-armed bandit learning approach is adopted to learn the behaviors of the users and a novel aggregation strategy is developed for residential heating,ventilation,and air conditioning(HVAC)to provide reliable secondary frequency *** with the conventional approach,the simulation results show that the risk-averse multiarmed bandit learning approach performs better in secondary frequency regulation with fewer users being selected and opting out of the ***,the proposed approach is more robust to random and changing behaviors of the users.

关键词： Heating ventilation and air conditioning(HVAC) load control multi-armed bandit online learning secondary frequency regulation

Probabilistic and Semantic Descriptions of Image Manifolds and Their Applications

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Tu, Peter Yang, Zhaoyuan Hartley, Richard Xu, Zhiwei Zhang, Jing Fu, Yiwei Campbell, Dylan Singh, Jaskirat Wang, Tianyu Computer Vision and Machine Learning Laboratory General Electric Research NiskayunaNY United States School of Computing College of Engineering Computing and Cybernetics Australian National University Canberra Australia

This paper begins with a description of methods for estimating probability density functions for images that reflects the observation that such data is usually constrained to lie in restricted regions of the high-dimensional image space — not every pattern of pixels is an image. It is common to say that images lie on a lower-dimensional manifold in the high-dimensional space. However, although images may lie on such lower-dimensional manifolds, it is not the case that all points on the manifold have an equal probability of being images. Images are unevenly distributed on the manifold, and our task is to devise ways to model this distribution as a probability distribution. In pursuing this goal, we consider generative models that are popular in AI and computer vision community. For our purposes, generative/probabilistic models should have the properties of 1) sample generation: it should be possible to sample from this distribution according to the modelled density function, and 2) probability computation: given a previously unseen sample from the dataset of interest, one should be able to compute the probability of the sample, at least up to a normalising constant. To this end, we investigate the use of methods such as normalising flow and diffusion models. We then show how semantic interpretations are used to describe points on the manifold. To achieve this, we consider an emergent language framework that makes use of variational encoders to produce a disentangled representation of points that reside on a given manifold. Trajectories between points on a manifold can then be described in terms of evolving semantic descriptions. In addition to describing the manifold in terms of density and semantic disentanglement, we also show that such probabilistic descriptions (bounded) can be used to improve semantic consistency by constructing defences against adversarial attacks. We evaluate our methods on CelebA and point samples for likelihood estimation with improved seman

关键词： Semantics