Xiangxiang Xu

Dr. Xiangxiang Xu

orcid

Research Focus & Metrics

Academic Career Length

13Years (2014-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
17 July 2019 in Proceedings of the AAAI Conference on Artificial Intelligence
Verified
Journal Article
Closed access
PDF file
LWLichen Wang
Lichen Wang
JWJiaxiang Wu
Jiaxiang Wu
SHShao-Lun Huang
Shao-Lun Huang
LZLizhong Zheng
Lizhong Zheng
XXXiangxiang Xu
Xiangxiang Xu
JHJunzhou Huang
Junzhou Huang
One primary focus in multimodal feature extraction is to find the representations of individual modalities that are maximally correlated. As a well-known measure of dependence, the Hirschfeld-Gebelein-Rényi (HGR) maximal correlation be-´ comes an appealing objective because of its operational meaning and desirable properties. However, the strict whitening constraints formalized in the HGR maximal correlation limit its application. To address this problem, this paper proposes Soft-HGR, a novel framework to extract informative features from multiple data modalities. Specifically, our framework prevents the “hard” whitening constraints, while simultaneously preserving the same feature geometry as in the HGR maximal correlation. The objective of Soft-HGR is straightforward, only involving two inner products, which guarantees the efficiency and stability in optimization. We further generalize the framework to handle more than two modalities and missing modalities. When labels are partially available, we enhance the discriminative power of the feature representations by making a semi-supervised adaptation. Empirical evaluation implies that our approach learns more informative feature mappings and is more efficient to optimize.

8 May 2017 in IET Generation, Transmission & Distribution
Verified
Journal Article
Open access
PDF file
TYTianyu Yang
Tianyu Yang
XXXiangxiang Xu
Xiangxiang Xu
QGQinglai Guo
Qinglai Guo
LZLin Zhang
Lin Zhang
HSHongbin Sun
Hongbin Sun
With the growing application of electric vehicles (EVs), it is of great significance to have a deep understanding of EV users driving and charging patterns for charging forecasting. However, the rapid growth scale of EV taxis with charging patterns that are closely coupled with human behaviours of temporal–spatial charging choices was not compatible with most previous coordinated strategies. Unlike the majority of existing approaches, a large volume of second‐level EV global positioning system (GPS) data was used to study the behaviour patterns of EV users. In practise, a mobile crowdsensing system that records GPS data and transmits information to the server was deployed in a fleet of electric‐taxi cabs in Shenzhen, China, making it possible to record the exact behaviour of each vehicle. Travelling and charging statuses of EVs were recorded and analysed into different characteristics of behaviour for each user. The load forecast methods proved to be more effective with more knowledge of both history and real‐time data.

8 October 2018
Verified
Conference Paper
Open access
PDF file
XCXinlei Chen
Xinlei Chen
XXXiangxiang Xu
Xiangxiang Xu
XLXinyu Liu
Xinyu Liu
SPShijia Pan
Shijia Pan
JHJiayou He
Jiayou He
PZPei Zhang
Pei Zhang

14 November 2016
Verified
Conference Paper
Open access
XXXiangxiang Xu
Xiangxiang Xu
XCXinlei Chen
Xinlei Chen
XLXinyu Liu
Xinyu Liu
HNHae Young Noh
Hae Young Noh
PZPei Zhang
Pei Zhang
LZLin Zhang
Lin Zhang

14 November 2016
Verified
Conference Paper
Open access
PDF file
XCXinlei Chen
Xinlei Chen
XXXiangxiang Xu
Xiangxiang Xu
XLXinyu Liu
Xinyu Liu
HNHae Young Noh
Hae Young Noh
LZLin Zhang
Lin Zhang
PZPei Zhang
Pei Zhang

3 February 2020 in IEEE Access
Verified
Journal Article
Open access
XXXiangxiang Xu
Xiangxiang Xu
SHShao-Lun Huang
Shao-Lun Huang
In this paper, we propose a novel regression analysis approach, called maximal correlation regression, by exploiting the ideas from the Hirschfeld-Gebelein-Rényi (HGR) maximal correlation. We show that in supervised learning problems, the optimal weights in maximal correlation regression can be expressed analytically with the relationships to the HGR maximal correlation functions, which reveals theoretical insights for our approach. In addition, we apply the maximal correlation regression to deep learning, in which efficient training algorithms are proposed for learning the weights in hidden layers. Furthermore, we illustrate that the maximal correlation regression is deeply connected to several existing approaches in information theory and machine learning, including the universal feature selection problem, linear discriminant analysis, and the softmax regression. Finally, experiments on real datasets demonstrate that our approach can obtain performance comparable to the widely used softmax regression based-method.