Wei Liu (刘伟)

Contacts:

Phone: +61 2 9514 3782

Email: wei.liu[at]uts.edu.au

Address: Room 11.07.307, Level 7, Building 11 (the Broadway building)PO Box 123, Broadway NSW 2007, AUSTRALIA


Short Biography:

I'm a Lecturer at the Advanced Analytics Institute within the School of Software, Faculty of Engineering and IT, the University of Technology, Sydney. 

Before joining UTS, I was a Research Fellow at the University of Melbourne working with Rao Kotagiri, James Bailey, and Chris Leckie, and then an Industry-focused Machine Learning Researcher at NICTA working with the transportation industry. I was educated at the University of Sydney where I obtained my PhD degree in Sanjay Chawla's group. During my PhD, I visited Microsoft Research Asia in Yu Zheng's research group.

I have been publishing papers in the following research areas (which are my major research interests): graph mining, causal inference, tensor factorization, multi-modal feature selection, game theory, data imbalance learning, spatio-temporal data mining, discrimination-aware learning, and anomaly (outlier) detection.


Publications (here is my Google Scholar page):

      Selected top-tier papers: 
  • Khoa Nguyen, Wei Liu, Fang Chen, Peter Runcie, et al.: Using Tensor Analysis for Damage Identification in Civil Structures. In proceedings of the 19th Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2015). Accepted as a Long Presentation -- acceptance rate 7%
    ERA ranking: A
  • Xinxin Jiang, Wei Liu, Longbing Cao, Guodong Long: Coupled Collaborative Filtering for Context-aware Recommendation. In Proceedings of the Twenty-Ninth AAAI Conference on Artificial Intelligence (AAAI 2015). 
    ERA ranking: A
  • Jingyu Shao, Junfu Yin, Wei Liu, Longbing Cao: Actionable Combined High Utility Itemset Mining. In Proceedings of the Twenty-Ninth AAAI Conference on Artificial Intelligence (AAAI 2015). 
    ERA ranking: A
  • Fei Wang, Wei Liu, Sanjay Chawla: On Sparse Feature Attacks in Adversarial Learning. In proceedings of the 2014 IEEE International Conference on Data Mining (ICDM 2014). (Acceptance rate: 19%)
    ERA ranking: A
  • Wei Liu, Dong Lee, and Rao Kotagiri: Significantly Improving General Supervised Learners by Using Local Information.  In proceedings of the 23rd ACM Conference on Information and Knowledge Management (CIKM 2014). (Acceptance rate: 17%)
    ERA ranking: A
  • Victor Chu, Wei Liu, Raymond Wong, Fang Chen, and Charles Perng: Traffic Analysis as a Service via a Unified Model. In proceedings of the 11th IEEE International Conference on Services Computing (IEEE SCC 2014). 
    ERA ranking: A
  • Victor Chu, Wei Liu, Raymond Wong, and Fang Chen: Causal Structure Discovery for Spatio-temporal Data. In proceedings of the 19th International Conference on Database Systems for Advanced Applications (DASFAA 2014). 
    ERA ranking: A
  • Jeffrey Chan, Wei Liu, James Bailey, Christopher Leckie, and Rao Kotagiri: Structure-aware Distance Measures for Comparing Clusterings in Graphs. In proceedings of the 18th Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2014). 
    ERA ranking: A

         Other papers

      • Ling Luo, Irena Koprinska and Wei Liu. Discrimination-Aware Classifiers for Student Performance Prediction. In Proceedings of the 8th International Conference on Educational Data Mining (EDM 2015).
      • ERA ranking: B 
      • Linsey Pang, Sanjay Chawla, Wei Liu and Yu Zheng: On Detection of Emerging Anomalous Traffic Patterns Using GPS Data. Data & Knowledge Engineering journal, vol. 87, pages 357-373, 2013. 
        ERA ranking: B
           

      Industry Engagement

      I have led several major research projects in partnership with government agencies and industrial organisations in Internet Security, Insurance, Capital Market, Transportation, and Infrastructure sectors. I developed cutting-edge data mining methods and software tools for the transport industry, which accurately identify causes of road incidents from dynamic traffic networks. I also designed advanced predictive analysis models for the problems of rare event prediction, fraud/intrusion detection, and spam filtering with time-evolving data distributions, which are validated and applied in different industrial organisations. Details of some of my research projects are in the below:

      •  “Traffic Watch for Transport Control Service”, industry partner: Transport Management Centre; May 2013 – June 2014.
      •  “Congestion Propagation and Hotspot Detection in Sydney CBD”, industry partner: NSW RMS; Aug – Dec 2013.
      •  “Data Fusion Technologies for Comprehensive Transport Data Analysis in Melbourne”, industry partner: VicRoads; Jun – Sep 2013.
      •  “Time of Arrival Estimations using HD Vehicle Trajectories”, industry partner: Tomtom. Jan 2013 – March 2013.
      •  “Early Detection of Road Traffic Incidents using Social Media”, industry partner: the Transport Management Centre; Oct – Dec 2012.
      •  “Causal Inference for Sequential Traffic Congestion", industry partner: Microsoft Research Asia; Nov 2010 – Mar 2011.
      •  “Abnormal Claim Detection from Worker’s Compensations”, industry partner: CGU Insurance; Mar 2010 – Jun 2011.
      •  “Data Integration for Cross-Market Capital Trading Systems”, industry partner: the SMARST Group (now purchased by Nasdaq), Jun 2008 – Dec 2009.


      Current Student Collaboration / Supervision:

      • Shoujin Wang, PhD student, on non-i.i.d. pattern mining 
      • Xinxin Jiang, PhD student, on context-aware recommendation systems
      • Ling Luo, PhD student, on causal inference for spatio-temporal data   
      • Fei Wang, PhD student, on game theory for machine learning  

      Completed Student Supervision / Collaboration

      • Victor Chu, PhD student, causal structure discovery from road traffic data  
      • Goce Ristanoski, PhD student, on time series regression (with Prof James Bailey) 
      • Mohammad Mazraehshahi, Master student, on soft-cut decision trees (with Prof Rao Kotagiri)
      • DongHawn Lee, Master student, on lazy learning methods (with Prof Rao Kotagiri) 
      • Jinheng Liu, Master student, ensembles for outlier detection
      • Emma Wang, Master student, feature creation for Bagging 
      • Chun Yuan Lee, Summer intern student, Causality Models for Understanding Road Traffic Changes
      • Amogh Sarda, Summer intern student, Incident impact analysis by earning historical road traffic data 
      • Han Zhou, Summer intern student, Mining social networks for early detection of events  
      • Zeyang Yu, Summer intern student, on Twitter-based road incident detection. 
      • Mingxuan Li, Summer intern student (winner of NICTA summer scholar prize), on mining dynamic road traffics.

       

      My Readings/Focus (some via library login)

        A* / A ranking venues (rankings are from arc.gov.au):

       Others: