Xun Wang

Xun Wang

Auto-generated profile, based on publicly available information

Xun Wang

Zhejiang Gongshang University, Hangzhou, China

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
24 December 2020 in IEEE Transactions on Knowledge and Data Engineering
Verified
Journal Article
Closed access
RSRenjie Sun
Renjie Sun
CCChen Chen
Chen Chen
XWXiaoyang Wang
Xiaoyang Wang
YZYing Zhang
Ying Zhang
XWXun Wang
Xun Wang
Community detection is one of the most fundamental problems in social network analysis, while most existing research focuses on unsigned graphs. In real applications, social networks involve not only positive relationships but also negative ones. It is important to exploit the signed information to identify more stable communities. In this paper, we propose a novel model, named stable k-core, to measure the stability of a community in signed graphs. The stable k-core model not only emphasizes user engagement, but also eliminates unstable structures. We show that the problem of finding the maximum stable k-core is NP-hard. To scale for large graphs, novel pruning strategies and searching methods are proposed. We conduct extensive experiments on 6 real-world signed networks to verify the efficiency and effectiveness of proposed model and techniques.

6 February 2023 in ACM Transactions on Multimedia Computing, Communications, and Applications
Verified
Journal Article
Open access
QZQi Zheng
Qi Zheng
JDJianfeng Dong
Jianfeng Dong
XQXiaoye Qu
Xiaoye Qu
XYXun Yang
Xun Yang
YWYabing Wang
Yabing Wang
XWXun Wang
Xun Wang
This article targets the task of language-based video moment localization. The language-based setting of this task allows for an open set of target activities, resulting in a large variation of the temporal lengths of video moments. Most existing methods prefer to first sample sufficient candidate moments with various temporal lengths, then match them with the given query to determine the target moment. However, candidate moments generated with a fixed temporal granularity may be suboptimal to handle the large variation in moment lengths. To this end, we propose a novel multi-stage Progressive Localization Network (PLN) that progressively localizes the target moment in a coarse-to-fine manner. Specifically, each stage of PLN has a localization branch and focuses on candidate moments that are generated with a specific temporal granularity. The temporal granularities of candidate moments are different across the stages. Moreover, we devise a conditional feature manipulation module and an upsampling connection to bridge the multiple localization branches. In this fashion, the later stages are able to absorb the previously learned information, thus facilitating the more fine-grained localization. Extensive experiments on three public datasets demonstrate the effectiveness of our proposed PLN for language-based moment localization, especially for localizing short moments in long videos.

26 October 2023
Verified
Conference Paper
Open access
SSShengkai Sun
Shengkai Sun
DLDaizong Liu
Daizong Liu
JDJianfeng Dong
Jianfeng Dong
XQXiaoye Qu
Xiaoye Qu
JGJunyu Gao
Junyu Gao
MWMeng Wang
Meng Wang

19 February 2024 in IEEE Transactions on Image Processing
Verified
Journal Article
Closed access
YWYabing Wang
Yabing Wang
SWShuhui Wang
Shuhui Wang
HLHao Luo
Hao Luo
JDJianfeng Dong
Jianfeng Dong
FWFan Wang
Fan Wang
MWMeng Wang
Meng Wang
Current research on cross-modal retrieval is mostly English-oriented, as the availability of a large number of English-oriented human-labeled vision-language corpora. In order to break the limit of non-English labeled data, cross-lingual cross-modal retrieval (CCR) has attracted increasing attention. Most CCR methods construct pseudo-parallel vision-language corpora via Machine Translation (MT) to achieve cross-lingual transfer. However, the translated sentences from MT are generally imperfect in describing the corresponding visual contents. Improperly assuming the pseudo-parallel data are correctly correlated will make the networks overfit to the noisy correspondence. Therefore, we propose Dual-view Curricular Optimal Transport (DCOT) to learn with noisy correspondence in CCR. In particular, we quantify the confidence of the sample pair correlation with optimal transport theory from both the cross-lingual and cross-modal views, and design dual-view curriculum learning to dynamically model the transportation costs according to the learning stage of the two views. Extensive experiments are conducted on two multilingual image-text datasets and one video-text dataset, and the results demonstrate the effectiveness and robustness of the proposed method. Besides, our proposed method also shows a good expansibility to cross-lingual image-text baselines and a decent generalization on out-of-domain data.

Verified
Journal Article
Open access
PDF file
JZJinming Zhang
Jinming Zhang
XWXun Wang
Xun Wang
LXLianrui Xu
Lianrui Xu
XZXin Zhang
Xin Zhang
Commonly used robot map models include occupancy grid maps, topological maps, and semantic maps. Among these, an occupancy grid map is mainly represented as a quadrilateral grid. This paper proposes an occupancy information grid for intelligent robots by exploiting the advantages of the occupancy grid map and spatial information grid. In terms of geometric structure, a regular hexagonal grid is used instead of a regular quadrilateral grid. In terms of attribute structure, the single obstacle attribute is replaced by the grid terrain characteristics, grid element attributes, and grid edge attributes. Thus, the occupancy information grid model is transformed into a new data structure describing the spatial environment, and it can be effectively applied to map construction and path planning of intelligent robots. For the map construction application of intelligent robots, this paper describes the basic process of laser sensor-based grid model construction. For the path planning application of intelligent robots, this paper extends the A* algorithm based on a regular hexagonal grid. Additionally, map construction and path planning applications for intelligent robots are experimentally verified. Several experimental results were obtained. First, the experimental results confirmed the theoretical conclusion that the minimum sampling density of the hexagonal structure was 13.4% lower than that of the quadrilateral structure. Second, the regular hexagonal grid is clearly more advantageous in describing environmental scenes, which can ameliorate the "undercompleteness" phenomenon. Third, there were large differences in the planning paths based on two types of grids, as shown by the fact that the distance of the planning paths obtained by the regular hexagonal grid was reduced by at least 10.8% and at most 15.6% compared with the regular quadrilateral grid.