Search

Your search keyword '"WANG Nan"' showing total 22,872 results
Start Over Author "WANG Nan"
22,872 results on '"WANG Nan"'

438. IMO$^3$: Interactive Multi-Objective Off-Policy Optimization

Author

Wang, Nan, Wang, Hongning, Karimzadehgan, Maryam, Kveton, Branislav, and Boutilier, Craig

Subjects
Computer Science - Machine Learning, Computer Science - Computational Engineering, Finance, and Science
Abstract

Most real-world optimization problems have multiple objectives. A system designer needs to find a policy that trades off these objectives to reach a desired operating point. This problem has been studied extensively in the setting of known objective functions. We consider a more practical but challenging setting of unknown objective functions. In industry, this problem is mostly approached with online A/B testing, which is often costly and inefficient. As an alternative, we propose interactive multi-objective off-policy optimization (IMO$^3$). The key idea in our approach is to interact with a system designer using policies evaluated in an off-policy fashion to uncover which policy maximizes her unknown utility function. We theoretically show that IMO$^3$ identifies a near-optimal policy with high probability, depending on the amount of feedback from the designer and training data for off-policy estimation. We demonstrate its effectiveness empirically on multiple multi-objective optimization problems.

Published
2022

439. Theoretical Limits of Joint Detection and Estimation for Radar Target

Author

Wang, Nan and Xu, Dazhuan

Subjects
Computer Science - Information Theory
Abstract

This paper proposes a joint detection and estimation (JDE) scheme based on mutual information for the radar work, whose goal is to choose the true one between target existent and target absence, and to estimate the unknown distance parameter when the target is existent. Inspired by the thoughts of Shannon information theory, the JDE system model is established in the presence of complex white Gaussian noise. We make several main contributions: (1) the equivalent JDE channel and the posterior probability density function are derived based on the priori statistical characteristic of the noise, target scattering and joint target parameter; (2) the performance of the JDE system is measured by the joint entropy deviation and the joint information that is defined as the mutual information between received signal and the joint target parameter; (3) the sampling a posterior probability and cascaded JDEers are proposed, and their performance is measured by the empirical joint entropy deviation the empirical joint information; (4) the joint theorem is proved that the joint information is the available limit of the overall performance, that is, the joint information is available, and the empirical joint information of any JDEer is no greater than the joint information; (5) the cascaded theorem is proved that the sum of empirical detection information and empirical estimation information can approximate the joint information, i.e., the performance limit of cascaded JDEer is available. Simulation results verify the correctness of the joint and the cascaded theorems, and show that the performance of the sampling a posterior probability JDEer is asymptotically optimal. Moreover, the performance of cascaded JDEer can approximate the system performance of JDE system., Comment: 18 pages, 7 figures

Published
2022

443. Symmetric inheritance of parental histones contributes to safeguarding the fate of mouse embryonic stem cells during differentiation

Author

Wen, Qing, Zhou, Jiaqi, Tian, Congcong, Li, Xinran, Song, Guibing, Gao, Yuan, Sun, Yaping, Ma, Chiyuan, Yao, Sitong, Liang, Xiaoyan, Kang, Xing, Wang, Nan, Yao, Yuan, Wang, Hongbao, Liang, Xiaohuan, Tang, Jialin, Offer, Steven M., Lei, Xiaohua, Yu, Chuanhe, Liu, Xiangyu, Liu, Zichuan, Wang, Zhiquan, and Gan, Haiyun

Published
2023
Full Text
View/download PDF

450. FedNI: Federated Graph Learning with Network Inpainting for Population-Based Disease Prediction

Author

Peng, Liang, Wang, Nan, Dvornek, Nicha, Zhu, Xiaofeng, and Li, Xiaoxiao

Subjects
Computer Science - Machine Learning, Computer Science - Social and Information Networks
Abstract

Graph Convolutional Neural Networks (GCNs) are widely used for graph analysis. Specifically, in medical applications, GCNs can be used for disease prediction on a population graph, where graph nodes represent individuals and edges represent individual similarities. However, GCNs rely on a vast amount of data, which is challenging to collect for a single medical institution. In addition, a critical challenge that most medical institutions continue to face is addressing disease prediction in isolation with incomplete data information. To address these issues, Federated Learning (FL) allows isolated local institutions to collaboratively train a global model without data sharing. In this work, we propose a framework, FedNI, to leverage network inpainting and inter-institutional data via FL. Specifically, we first federatively train missing node and edge predictor using a graph generative adversarial network (GAN) to complete the missing information of local networks. Then we train a global GCN node classifier across institutions using a federated graph learning platform. The novel design enables us to build more accurate machine learning models by leveraging federated learning and also graph learning approaches. We demonstrate that our federated model outperforms local and baseline FL methods with significant margins on two public neuroimaging datasets.

Published
2021
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results