Welcome! My name is Wenjie Hu and I am currently a R&D engineer of Alibaba Cloud. My research interests lie in mining deep knowledge from large-scale data, which focus on Time Series Modeling and Graph Embedding. I obtained my master degree from Zhejiang University in 2020, advised by Yang Yang in Digital media Computing & Design (DCD) Lab. I also fortunately have worked with Xiang Ren from University of Southern California. For more detailed personal information, please refer to my resume.
Our research paper, EvoNet, the second work of revisiting time series modeling via graph, has been accepted by WSDM 2021 (Oral 11.4%).
I served as a PC member at SMP 2020.
I am awarded the Excellent Graduate of Zhejiang University, along with the Graduate Scholarship (Top 3%).
The work of detecting electricity theft, cooperated with State Grid of Zhejiang, China, has been accepted by WWW 2020 (Oral 8.5%).
Our research paper, Time2Graph, a work of revisiting time series modeling via graph, has been accepted by AAAI 2020 .
I am awarded the Vmware Excellent Student Scholarship (Top 5%).
Time series modeling has attracted extensive research efforts; however, achieving both reliable efficiency and interpretability from a unified model still remains a challenging problem.
Our recent work proposes to model time series from the perspective of graphs. More specifically, we aim to capture the representative patterns (or referred to as states) and their transitions underpinning the observed time series, and describe how these factors affect the time series evolution. To achieve this, we respectively propose the shapelet based method (Time2Graph, Cheng et al., AAAI'20) and a dynamic graph neural network based model (EvoNet, Hu et al., WSDM'21). Our proposed methods not only achieves clear improvements comparing with state-of-the-art baselines in many tasks, but also provide valuable insights towards explaining the results of prediction results.
Our work has been applied in real-world scenarios, such as anomaly detection of time series, as a common service of Alibaba Cloud, collaborated with SLS, Alibaba Cloud, Alibaba Group. Currently our algorithm have served hundreds of enterprise users. ( Hu et al., WSDM'21, Cheng et al., AAAI'20) [Project Homepage]
Electricity theft, the behavior that involves users conducting illegal operations on electrical meters to avoid individual electricity bills, is a common phenomenon in the developing countries. Considering its harmfulness to both power grids and the public, substantial effort has been expended to prevent or detect these behaviors.
To explore the patterns of electricity theft hidden in the large-scale users, we conduct a deep study on the multi-source data: in addition to users’ electricity usage records, we analyze user behaviors by means of regional factors (non-technical loss) and climatic factors (temperature) in the corresponding transformer area. Our results unearth several interesting patterns: electricity thieves are likely to consume much more electrical power than normal users, especially under extremely high or low temperatures. Motivated by these empirical observations, a novel hierarchical framework is designed for identifying electricity thieves. (Hu et al., WWW'20)
The followings are some of my works, including the public address of github repository
When a reporter asked Bryant why so successful? Bryant asked back: 'Do you know Los Angeles morning four-point looks like?
Discover the potential information, release your commercial power.
Design effective models and incubate industry solutions.
Develop effective system, improve the stability and intelligence.
It's great to have friends coming from afar.