UTS site search

Dr Wei Liu


Dr Wei Liu is a Data Science Research Program Leader and a Lecturer at the Advanced Analytics Institute, Faculty of Engineering and IT, the University of Technology Sydney (UTS). Before joining UTS, he was a Data Mining Research Fellow at the University of Melbourne, and then an Industry-focused Machine Learning Researcher and Project Manager working in the transportation industry at National ICT Australia (NICTA). He obtained his PhD degree in Data Mining Research from the University of Sydney (USYD).

His research outputs are mostly published in journals and conferences that are ranked at A* and A (i.e., top-prestige) by ARC ERA 2010 ranking and by Core academic ranking. He has received 3 Best Paper Awards.

Dr Liu has been engaged in industry-driven data analytics research that makes real-world impact. He has led a number of significant research projects funded by government agencies and industrial organisations, spanning internet security, insurance, trading, transportation, and infrastructure sectors. He has developed advanced data mining models and software tools for the transport industry, which accurately identify causes of road incidents. He has also designed cutting-edge predictive models for problems including rare event prediction, fraud/intrusion detection, emerging trends detection, etc. Details of some of his research projects are in the below:

  • "Advanced Data Analytics Platforms without Data", industry partner: National ICT Australia; March 2016 – December 2018.
  • "Data Analytics Models for Stock Market Surveillance", industry partner: NASDAQ OMX; March 2016 - December 2018.
  • "Analytics Model to Support Strategic Planning in a Regulatory Environment", industry partner: NSW Fair Trading; April - July 2015.
  • "Transport Data Science and Advanced Analytics", industry partner: National ICT Australia; July 2015 – June 2017.
  • “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.

Image of Wei Liu
Lecturer, Faculty of Engineering & Information Technology
Core Member, AAI - Advanced Analytics Institute
Member, Institute of Electrical and Electronics Engineers
Member, Association for Computing Machinery
+61 2 9514 3782

Research Interests

Main Research Interests:

  • Graph mining, dynamic network analysis, tensor factorization
  • Causal inference, Granger causality
  • Game theoretical modeling, adversarial learning 
  • Data imbalance, cost-sensitive learning
  • Anomaly (outlier) detection
Data Mining and Knowledge Discovery; Data Analytics.


Chen, Q., Hu, L., Xu, J., Liu, W. & cao, L. 2015, 'Document Similarity Analysis via Involving Both Explicit and Implicit Semantic Couplings', 2015 International Conference on Data Science and Advanced Analytics, Paris.
View/Download from: UTS OPUS
Nguyen, H., Liu, W., Rivera, P. & Chen, F. 2016, 'TrafficWatch: Real-time traffic incident detection and monitoring using social media', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Pacific Asia Knowledge Discovery and Data Mining Conference (PAKDD) 2016, New Zealand, pp. 540-551.
View/Download from: Publisher's site
© Springer International Publishing Switzerland 2016. Social media has become a valuable source of real-time information. Transport Management Centre (TMC) in Australian state government of New South Wales has been collaborating with us to develop TrafficWatch, a system that leverages Twitter as a channel for transport network monitoring, incident and event managements. This system utilises advanced web technologies and state-of-the-art machine learning algorithms. The crawled tweets are first filtered to show incidents in Australia, and then divided into different groups by online clustering and classification algorithms. Findings from the use of TrafficWatch at TMC demonstrated that it has strong potential to report incidents earlier than other data sources, as well as identifying unreported incidents. TrafficWatch also shows its advantages in improving TMC's network monitoring capabilities to assess network impacts of incidents and events.
Do, D. & Liu, L. 2016, 'ASTEN: an Accurate and Scalable Approach to Coupled Tensor Factorization', the International Joint Conference in Neural Networks, the International Joint Conference in Neural Networks, IEEE, Vancouver, Canada, pp. 99-106.
View/Download from: UTS OPUS or Publisher's site
Coupled Tensor Factorization (CTF) has become one of the most popular methods for joint analysis of high dimensional data generated from multiple sources. The goal of CTF is to factorize correlated datasets into latent factors efficiently. This research was taken with a particular goal of improving the accuracy of CTF. It is important to optimize the factorization of each single tensor of the coupled tensors. To achieve this, we introduce ASTEN, an Accurate and Scalable Tensor factorization method, where the objective function is optimized with respect to every single tensor and matrix. Differing from algorithms with a traditional objective function which forces shared modes among tensors to have identical factors, ASTEN enables each tensor to have its own discriminative factor on the shared mode and thus is capable of finding the accurate approximation of every tensor. Furthermore, to make it highly scalable in handling big data, we design it to be fully distributed and scalable with respect to the number of tensors, their dimensions, their sizes and the number of data partitions. In addition, we provide our theoretical proof and experimental evidence that our algorithm converges to an optimum. Experiments on both real and synthetic datasets demonstrate that our proposed ASTEN outperforms alternative existing algorithms.
Rashidi, L., Kan, A., Bailey, J., Chan, J., Leckie, C., Liu, W., Rajasegarar, S. & Ramamohanarao, K. 2016, 'Node re-ordering as a means of anomaly detection in time-evolving graphs', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Joint European Conference on Machine Learning and Knowledge Discovery in Databases, Springer, Riva del Garda, Italy, pp. 162-178.
View/Download from: UTS OPUS or Publisher's site
© Springer International Publishing AG 2016.Anomaly detection is a vital task for maintaining and improving any dynamic system. In this paper, we address the problem of anomaly detection in time-evolving graphs, where graphs are a natural representation for data in many types of applications. A key challenge in this context is how to process large volumes of streaming graphs. We propose a pre-processing step before running any further analysis on the data, where we permute the rows and columns of the adjacency matrix. This pre-processing step expedites graph mining techniques such as anomaly detection, PageRank, or graph coloring. In this paper, we focus on detecting anomalies in a sequence of graphs based on rank correlations of the reordered nodes. The merits of our approach lie in its simplicity and resilience to challenges such as unsupervised input, large volumes and high velocities of data. We evaluate the scalability and accuracy of our method on real graphs, where our method facilitates graph processing while producing more deterministic orderings. We show that the proposed approach is capable of revealing anomalies in a more efficient manner based on node rankings. Furthermore, our method can produce visual representations of graphs that are useful for graph compression.
Braytee, A., liu & kennedy 2016, 'A Cost-Sensitive Learning Strategy for Feature Extraction from Imbalanced Data', Springer International Publishing, International Conference on Neural Information Processing, Springer International Publishing, Kyoto, Japan, pp. 78-86.
View/Download from: UTS OPUS or Publisher's site
In this paper, novel cost-sensitive principal component analysis (CSPCA) and cost-sensitive non-negative matrix factorization (CSNMF) methods are proposed for handling the problem of feature extraction from imbalanced data. The presence of highly imbalanced data misleads existing feature extraction techniques to produce biased features, which results in poor classification performance especially for the minor class problem. To solve this problem, we propose a cost-sensitive learning strategy for feature extraction techniques that uses the imbalance ratio of classes to discount the majority samples. This strategy is adapted to the popular feature extraction methods such as PCA and NMF. The main advantage of the proposed methods is that they are able to lessen the inherent bias of the extracted features to the majority class in existing PCA and NMF algorithms. Experiments on twelve public datasets with different levels of imbalance ratios show that the proposed methods outperformed the state-of-the-art methods on multiple classifiers.
Do, D., Pham, Liu & Ramamohanarao 2016, 'WTEN: An Advanced Coupled Tensor Factorization Strategy for Learning from Imbalanced Data', Web Information Systems Engineering – WISE 2016, the 17th International Conference on Web Information Systems Engineering (WISE), Springer International Publishing, Shanghai, China, pp. 537-552.
View/Download from: UTS OPUS or Publisher's site
Liu, W. 2016, 'Factorization of multiple tensors for supervised feature extraction', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), International Conference on Neural Information Processing, Springer, Kyoto, Japan, pp. 406-414.
View/Download from: UTS OPUS or Publisher's site
© Springer International Publishing AG 2016.Tensors are effective representations for complex and timevarying networks. The factorization of a tensor provides a high-quality low-rank compact basis for each dimension of the tensor, which facilitates the interpretation of important structures of the represented data. Many existing tensor factorization (TF) methods assume there is one tensor that needs to be decomposed to low-rank factors. However in practice, data are usually generated from different time periods or by different class labels, which are represented by a sequence of multiple tensors associated with different labels. When one needs to analyse and compare multiple tensors, existing TF methods are unsuitable for discovering all potentially useful patterns, as they usually fail to discover either common or unique factors among the tensors: (1) if each tensor is factorized separately, the factor matrices will fail to explicitly capture the common information shared by different tensors, and (2) if tensors are concatenated together to form a larger 'overall tensor and then factorize this concatenated tensor, the intrinsic unique subspaces that are specific to each tensor will be lost. The cause of such an issue is mainly from the fact that existing tensor factorization methods handle data observations in an unsupervised way, considering only features but not labels of the data. To tackle this problem, we design a novel probabilistic tensor factorization model that takes both features and class labels of tensors into account, and produces informative common and unique factors of all tensors simultaneously. Experiment results on feature extraction in classification problems demonstrate the effectiveness of the factors discovered by our method.
Braytee, A., Catchpoole, D.R., Kennedy, P.J. & Liu, W. 2016, 'Balanced Supervised Non-Negative Matrix Factorization for Childhood Leukaemia Patients', CIKM '16 Proceedings of the 25th ACM International on Conference on Information and Knowledge Management, ACM International on Conference on Information and Knowledge Management, ACM, Indianapolis, Indiana, USA.
View/Download from: Publisher's site
Supervised feature extraction methods have received considerable attention in the data mining community due to their capability to improve the classification performance of the unsupervised dimensionality reduction methods. With increasing dimensionality, several methods based on supervised feature extraction are proposed to achieve a feature ranking especially on microarray gene expression data. This paper proposes a method with twofold objectives: it implements a balanced supervised non-negative matrix factorization (BSNMF) to handle the class imbalance problem in supervised non-negative matrix factorization techniques. Furthermore, it proposes an accurate gene ranking method based on our proposed BSNMF for microarray gene expression datasets. To the best of our knowledge, this is the first work to handle the class imbalance problem in supervised feature extraction methods. This work is part of a Human Genome project at The Children's Hospital at Westmead (TB-CHW), Australia. Our experiments indicate that the factorized components using supervised feature extraction approach have more classification capability than the unsupervised one, but it drastically fails at the presence of class imbalance problem. Our proposed method outperforms the state-of-the-art methods and shows promise in overcoming this concern.
Cheema, P., Khoa, N.L.D., Alamdari, M.M., Liu, W., Wang, Y., Chen, F. & Runcie, P. 2016, 'On Structural Health Monitoring Using Tensor Analysis and Support Vector Machine with Artificial Negative Data', CIKM '16 Proceedings of the 25th ACM International on Conference on Information and Knowledge Management, ACM International on Conference on Information and Knowledge Management, ACM, Indianapolis, Indiana, USA, pp. 1813-1822.
View/Download from: Publisher's site
Structural health monitoring is a condition-based technology to monitor infrastructure using sensing systems. Since we usually only have data associated with the healthy state of a structure, one-class approaches are more practical. However, tuning the parameters for one-class techniques (like one-class Support Vector Machines) still remains a relatively open and difficult problem. Moreover, in structural health monitoring, data are usually multi-way, highly redundant and correlated, which a matrix-based two-way approach cannot capture all these relationships and correlations together. Tensor analysis allows us to analyse the multi-way vibration data at the same time. In our approach, we propose the use of tensor learning and support vector machines with artificial negative data generated by density estimation techniques for damage detection, localization and estimation in a one-class manner. The artificial negative data can help tuning SVM parameters and calibrating probabilistic outputs, which is not possible to do with one-class SVM. The proposed method shows promising results using data from laboratory-based structures and also with data collected from the Sydney Harbour Bridge, one of the most iconic structures in Australia. The method works better than the one-class approach and the approach without using tensor analysis.
Wang, S., Liu, W., Wu, J., Cao, L., Meng, Q. & Kennedy, P.J. 2016, 'Training deep neural networks on imbalanced data sets', Proceedings of the International Joint Conference on Neural Networks, International Joint Conference on Neural Networks (IJCNN), IEEE, Vancouver, Canada, pp. 4368-4374.
View/Download from: UTS OPUS or Publisher's site
© 2016 IEEE.Deep learning has become increasingly popular in both academic and industrial areas in the past years. Various domains including pattern recognition, computer vision, and natural language processing have witnessed the great power of deep networks. However, current studies on deep learning mainly focus on data sets with balanced class labels, while its performance on imbalanced data is not well examined. Imbalanced data sets exist widely in real world and they have been providing great challenges for classification tasks. In this paper, we focus on the problem of classification using deep network on imbalanced data sets. Specifically, a novel loss function called mean false error together with its improved version mean squared false error are proposed for the training of deep networks on imbalanced data sets. The proposed method can effectively capture classification errors from both majority class and minority class equally. Experiments and comparisons demonstrate the superiority of the proposed approach compared with conventional methods in classifying imbalanced data sets on deep neural networks.
Jiang, X., Liu, W., Cao, L. & Long, G. 2015, 'Coupled Collaborative Filtering for Context-aware Recommendation', AAAI Publications, Twenty-Ninth AAAI Conference on Artificial Intelligence, Student Abstracts, AAAI 2015, AAAI, Austin Texas, USA, pp. 4172-4173.
View/Download from: UTS OPUS
Context-aware features have been widely recognized as important factors in recommender systems. However, as a major technique in recommender systems, traditional Collaborative Filtering (CF) does not provide a straight-forward way of integrating the context-aware information into personal recommendation. We propose a Coupled Collaborative Filtering (CCF) model to measure the contextual information and use it to improve recommendations. In the proposed approach, coupled similarity computation is designed to be calculated by interitem, intra-context and inter-context interactions among item, user and context-ware factors. Experiments based on different types of CF models demonstrate the effectiveness of our design.
Shao, J., Yin, J., Liu, W. & Cao, L. 2015, 'Mining Actionable Combined Patterns of High Utility and Frequency', Proceedings of the IEEE International Conference on Data Science and Advanced Analytics, IEEE International Conference on Data Science and Advanced Analytics, IEEE, Paris, pp. 1-10.
View/Download from: Publisher's site
In recent years, the importance of identifying actionable patterns has become increasingly recognized so that decision-support actions can be inspired by the resultant patterns. A typical shift is on identifying high utility rather than highly frequent patterns. Accordingly, High Utility Itemset (HUI) Mining methods have become quite popular as well as faster and more reliable than before. However, the current research focus has been on improving the efficiency while the coupling relationships between items are ignored. It is important to study item and itemset couplings inbuilt in the data. For example, the utility of one itemset might be lower than user-specified threshold until one additional itemset takes part in; and vice versa, an item's utility might be high until another one joins in. In this way, even though some absolutely high utility itemsets can be discovered, sometimes it is easily to find out that quite a lot of redundant itemsets sharing the same item are mined (e.g., if the utility of a diamond is high enough, all its supersets are proved to be HUIs). Such itemsets are not actionable, and sellers cannot make higher profit if marketing strategies are created on top of such findings. To this end, here we introduce a new framework for mining actionable high utility association rules, called Combined Utility-Association Rules (CUAR), which aims to find high utility and strong association of itemset combinations incorporating item/itemset relations. The algorithm is proved to be efficient per experimental outcomes on both real and synthetic datasets.
Luo, L., Liu, W., Koprinska, I. & Chen, F. 2015, 'Discovering causal structures from time series data via enhanced granger causality', AI 2015: Advances in Artificial Intelligence (LNCS), 28th Australasian Joint Conference on Artificial Intelligence, Springer, Canberra, Australia, pp. 365-378.
View/Download from: Publisher's site
© Springer International Publishing Switzerland 2015. Granger causality has been applied to explore predictive causal relations among multiple time series in various fields. However, the existence of nonstationary distributional changes among the time series variables poses significant challenges. By analyzing a real dataset, we observe that factors such as noise, distribution changes and shifts increase the complexity of the modelling, and large errors often occur when the underlying distribution shifts with time. Motivated by this challenge, we propose a new regression model for causal structure discovery – a Linear Model with Weighted Distribution Shift (linear WDS), which improves the prediction accuracy of the Granger causality model by taking into account the weights of the distribution-shift samples and by optimizing a quadratic-mean based objective function. The linear WDS is integrated in the Granger causality model to improve the inference of the predictive causal structure. The performance of the enhanced Granger causality model is evaluated on synthetic datasets and real traffic datasets, and the proposed model is compared with three different regression-based Granger causality models (standard linear regression, robust regression and quadratic-mean-based regression). The results show that the enhanced Granger causality model outperforms the other models especially when there are distribution shifts in the data.
Shao, J., Yin, J., Liu, W. & Cao, L. 2015, 'Actionable Combined High Utility Itemset Mining', AAAI'15 Proceedings of the Twenty-Ninth AAAI Conference on Artificial Intelligence, AAAI Conference on Artificial Intelligence Pages, pp. 4206-4207.
Luo, L., Liu, W., Koprinska, I. & Chen, F. 2015, 'Discrimination-aware association rule mining for unbiased data analytics', Big Data Analytics and Knowledge Discovery: 17th International Conference, DaWaK 2015, Valencia, Spain, September 1-4, 2015, Proceedings, International Conference on Big Data Analytics and Knowledge Discovery, Springer International Publishing, Valencia; Spain, pp. 108-120.
View/Download from: Publisher's site
A discriminatory dataset refers to a dataset with undesirable correlation between sensitive attributes and the class label, which often leads to biased decision making in data analytics processes. This paper investigates how to build discrimination-aware models even when the available training set is intrinsically discriminating based on some sensitive attributes, such as race, gender or personal status. We propose a new classification method called Discrimination-Aware Association Rule classifier (DAAR), which integrates a new discrimination-aware measure and an association rule mining algorithm. We evaluate the performance of DAAR on three real datasets from different domains and compare it with two non-discrimination-aware classifiers (a standard association rule classification algorithm and the state-of-the-art association rule algorithm SPARCCC), and also with a recently proposed discrimination-aware decision tree method. The results show that DAAR is able to effectively filter out the discriminatory rules and decrease the discrimination on all datasets with insignificant impact on the predictive accuracy.
Khoa, N.L.D., Zhang, B., Wang, Y., Liu, W., Chen, F., Mustapha, S. & Runcie, P. 2015, 'On Damage Identification in Civil Structures Using Tensor Analysis', Advances in Knowledge Discovery and Data Mining: 19th Pacific-Asia Conference Proceedings, Part 1, Pacific Asia Conference on Knowledge Discovery and Data Mining (PAKDD), Springer International Publishing, Ho Chi Minh City, Vietnam, pp. 459-471.
View/Download from: Publisher's site
Wang, F., Liu, W. & Chawla, S. 2015, 'On Sparse Feature Attacks in Adversarial Learning', Proceedings - IEEE International Conference on Data Mining, ICDM, IEEE International Conference on Data Mining, IEEE, Shenzhen; China, pp. 1013-1018.
View/Download from: Publisher's site
Adversarial learning is the study of machine learning techniques deployed in non-benign environments. Example applications include classifications for detecting spam email, network intrusion detection and credit card scoring. In fact as the gamut of application domains of machine learning grows, the possibility and opportunity for adversarial behavior will only increase. Till now, the standard assumption about modeling adversarial behavior has been to empower an adversary to change all features of the classifier sat will. The adversary pays a cost proportional to the size of 'attack'. We refer to this form of adversarial behavior as a dense feature attack. However, the aim of an adversary is not just to subvert a classifier but carry out data transformation in a way such that spam continues to appear like spam to the user as much as possible. We demonstrate that an adversary achieves this objective by carrying out a sparse feature attack. We design an algorithm to show how a classifier should be designed to be robust against sparse adversarial attacks. Our main insight is that sparse feature attacks are best defended by designing classifiers which use l1 regularizers.

Journal articles

Nguyen, H., Liu, W. & Chen, F. 2016, 'Discovering Congestion Propagation Patterns in Spatio-Temporal Traffic Data', IEEE Transactions on Big Data, pp. 1-1.
View/Download from: Publisher's site
Pang, L.X., Chawla, S., Liu, W. & Zheng, Y. 2013, 'On detection of emerging anomalous traffic patterns using GPS data', Data & Knowledge Engineering, vol. 87, pp. 357-373.
View/Download from: Publisher's site
Liu, W. & Chawla, S. 2010, 'Mining adversarial patterns via regularized loss minimization', Machine Learning, vol. 81, no. 1, pp. 69-83.
View/Download from: Publisher's site
Liu, W. & Huang, J., 'Adaptive Leader-Following Consensus for a Class of Higher-Order Nonlinear Multi-Agent Systems with Directed Switching Networks', Automatica, vol. 79, pp. 84-92.
View/Download from: Publisher's site
In this paper, we study the leader-following consensus problem for a class of uncertain nonlinear multi-agent systems under jointly connected directed switching networks. The uncertainty includes constant unbounded parameters and external disturbances. We first extend the recent result on the adaptive distributed observer from global asymptotical convergence to global exponential convergence. Then, by integrating the conventional adaptive control technique with the adaptive distributed observer, we present our solution by a distributed adaptive state feedback control law. Our result is illustrated by the leader-following consensus problem for a group of van der Pol oscillators.