Jun Xu

Jun Xu

Auto-generated profile, based on publicly available information

Jun Xu

Harbin Institute of Technology

Research Focus & Metrics

Academic Career Length

16Years (2011-Present)

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

Representative Publications

View All Publications
7 November 2022 in IEEE Transactions on Circuits and Systems for Video Technology
Verified
Journal Article
Closed access
ZLZejun Liu
Zejun Liu
FCFanglin Chen
Fanglin Chen
JXJun Xu
Jun Xu
WPWenjie Pei
Wenjie Pei
GLGuangming Lu
Guangming Lu
Cross-modal image-text retrieval is an important area of Vision-and-Language task that models the similarity of image-text pairs by embedding features into a shared space for alignment. To bridge the heterogeneous gap between the two modalities, current approaches achieve inter-modal alignment and intra-modal semantic relationship modeling through complex weighted combinations between items. In the intra-modal association and inter-modal interaction processes, the higher-weight items have a higher contribution to the global semantics. However, the same item always produces different contributions in the two processes, since most traditional approaches only focus on the alignment. This usually results in semantic changes and misalignment. To address this issue, this paper proposes Cross-modal Semantic Importance Consistency (CSIC) which achieves invariance in the semantic of items during aligning. The proposed technique measures the semantic importance of items obtained from intra-modal and inter-modal self-attention and learns a more reasonable representation vector by inter-calibrating the importance distribution to improve performance. We conducted extensive experiments on the Flickr30K and MS COCO datasets. The results show that our approach can significantly improve retrieval performance, proving the proposed approach’s superiority and rationality.

1 June 2020 in Automatica
Verified
Journal Article
Closed access
JXJun Xu
Jun Xu
QTQinghua Tao
Qinghua Tao
ZLZhen Li
Zhen Li
XXXiangming Xi
Xiangming Xi
JSJohan Suykens
Johan Suykens
SWShuning Wang
Shuning Wang

1 December 2012 in Journal of Process Control
Verified
Journal Article
Closed access
JXJun Xu
Jun Xu
XHXiaolin Huang
Xiaolin Huang
XMXiaomu Mu
Xiaomu Mu
SWShuning Wang
Shuning Wang

5 June 2023 in IEEE Transactions on Automation Science and Engineering
Verified
Journal Article
Closed access
HLHongyi Li
Hongyi Li
JXJun Xu
Jun Xu
QZQianchuan Zhao
Qianchuan Zhao
SWSixin Wang
Sixin Wang
Energy shortage is a challenge for many countries, and building energy consumption accounts for a considerable proportion of global energy consumption. The main work of this paper is to optimize the energy consumption of heating, ventilating, and air conditioning (HVAC) systems in buildings based on economic model predictive control (EMPC). The cost in EMPC design includes energy consumption and predicted mean vote (PMV), which is an index that evaluates the thermal comfort of indoor occupants. In order to model the nonlinearity of the PMV index, we propose a lattice piecewise linear (PWL) approximation, which has high approximation precision and facilitates the resulting optimization problem, which is basically a piecewise quadratic programming problem. For the piecewise quadratic programming, we propose a descent algorithm that converges quickly and scales well with the length of the prediction horizon in the EMPC problem. The experimental results demonstrate that the proposed method saves 19.78% of the electricity cost compared to the conventional control strategy and significantly increases indoor comfort. Note to Practitioners — The motivation of this article is to provide a control strategy to reduce building energy consumption and ensure indoor thermal comfort. In most of the existing methods for air conditioning temperature control, the occupants’ comfort hasn’t been considered. In this paper, thermal comfort is described by the PMV index, which is basically nonlinear. In order to model the thermal comfort more accurately, in this paper, the PMV index is approximated piecewise linearly in order to meet the requirements of accuracy and computational efficiency. The resulting optimization problem is not hard to solve, and we provide an efficient algorithm for solving this optimization problem. Preliminary simulation experiments demonstrate that this approach is practical, i.e., it achieves energy reduction and ensures thermal comfort. Our strategy, however, has not yet been deployed in real buildings. In future research, we will propose similar techniques for large-scale systems in order to solve energy optimization problems containing multiple thermal zones and realize the proposed technique in real buildings.

31 January 2022 in IEEE Transactions on Intelligent Transportation Systems
Verified
Journal Article
Open access
QTQinghua Tao
Qinghua Tao
ZLZhen Li
Zhen Li
JXJun Xu
Jun Xu
SLShu Lin
Shu Lin
BSBart De Schutter
Bart De Schutter
JSJohan Suykens
Johan Suykens
Traffic flow (TF) prediction is an important and yet a challenging task in transportation systems, since the TF involves high nonlinearities and is affected by many elements. Recently, neural networks have attracted much attention for TF prediction, but they are commonly black boxes with complex architectures and difficult to be interpreted, e.g., the contributions of specific traffic elements are not explicit, hardly providing informative guidance. In this paper, we aim at addressing more interpretable short-term TF prediction with joint consideration to high accuracy, and thus introduces a pragmatic method by applying the efficient hinging hyperplanes neural network (EHHNN) simply built upon sparse neuron connections. In the proposed method, different traffic factors are incorporated into the inputs, including their spatial-temporal information. Besides the pursuit of accuracy, we further extend the ANOVA decomposition of EHHNNs to the interpretation analysis with specifications to traffic data, in which the contributions concerning specific traffic variables are detected quantitatively. As such, the proposed method firstly applies the EHHNN to filter out more important traffic variables for dimensionality reduction while maintaining accurate prediction. Then, variable interpretation analysis is performed from different perspectives, e.g. to quantitatively investigate the influence of traffic factors and also their spatial-temporal impacts. Therefore, a predictor and an analyzing tool can both be attained for the TF by exerting the flexibility and extending the interpretability of EHHNNs, which is promising to provide informative guidance to future traffic control. Numerical experiments verify the effectiveness and potential of the proposed method in TF prediction and analysis.

1 December 2011
Verified
Conference Paper
Closed access
YNYun Niu
Yun Niu
LWLiji Wu
Liji Wu
LWLi Wang
Li Wang
XZXiangmin Zhang
Xiangmin Zhang
JXJun Xu
Jun Xu
A configurable IPSec processor for a high performance in-line network security processor that integrates two embedded 32-bit CPU cores, and an IPSec protocol processor on a SoC is presented. The IPSec processor can implement the transport/tunnel mode AH and ESP protocol of the IPSec, and support AES-128/192/256, HMAC-SHA-1 algorithm. The number of AH, ESP, AES, HMAC-SHA-1 IP-cores in the design can be configured for different use such as 10 Gigabit Ethernet and Gigabit Ethernet, even for the next generation 40/100G Network. Low power is also considered in the design. In the IPSec processor, crossbar switch architecture for multi-core data transfer is adopted. With four parallel AH, ESP, AES, HMAC-SHA-1 IP-cores separately connected to an 8x8 crossbar switch in the IPSec processor, a throughput of 1.5Gbps at 200MHz is achieved and hardware verification is implemented by FPGA. By simulation, the IPSec protocol operation can achieve 10Gbps wire speed with 32 IPSec protocol IP-cores and cryptographic IP-cores configured in the IPSec processor.