Lan Du

Faculty of Information Technology

Room 142, Building 63 - 25 Exhibition Walk

Monash University, Clayton, VIC 3800

Tel: +61 3 9905 9971

Email: given_name.surname@monash.edu

Intro

Dr. Lan Du is a lecturer (equivalent to assistant professor) at Monash University, and affiliated with the machine learning group in Centre of Data Science. He is a Teaching and Research Academic (a core faculty member) in the Faculty of IT, Monash Clayton Campus. He is also a deputy course director of master of data sceince degree at Monash.

He is interested in statistical modelling and learning from discrete data, such as texts. His research area broadly covers statistical machine leaning, natural language processing, data mining, and social network analysis.

 

Research Interests

  • Topic modelling: He is interested in developing advanced topic models that can take into account diverse information or features exhibited by free language text, for example, various linguistic feature (either semantic or syntactic), temporal information, labels, word embeddings, etc.
  • Non-parametric Bayesian methods: He has broad interests in the study of nonparametric Bayesian methods, e.g., Pitman-Yor process, Indian Buffet process, Gamma process, etc.
  • Probsblistic Tensor factorization for social network analysis, relation extraction and relational learning. He is particuarly interested in building various auxiliary information into the factorization process.
  • Statistical ranking models: Label ranking is an important task in, for example, preference learning. The existing ranking methods include but not limited to the pair-wised ranking models, distance based ranking models and multistage ranking models. However, if the number of items to be ranked is countably infinite, it is unfeasible for each ranker to rank all the items. He is interested in both partial ranking models and infinite ranking models, and applying them to, for example, topic segmentation.

Research highlights

  • "Dirichlet Belief Networks as Structured Topic Prior" co-authored with H. Zhao, W. Buntine, and M. Zhou, is accepted for publication in NIPS 2018. (NEW!)
  • "Inter and Intra Topic Structure Learning with Word Embeddings” co-authored with He Zhao, Wray Buntine and Mingyuan Zhou, is accepted for publication in ICML 2018. (NEW!)
  • "Leveraging Label Category Relationships in Multi-class Crowdsourcing" co-authored with Yuan Jin, Ye Zhu, and Mark Carman is accepted for publication in PAKDD 2018. (NEW!)
  • "Leveraging external information in topic modelling" co-authored with He Zhao, Wray Buntine and Gang Liu is published in KAIS 2018.(NEW!)
  • “Bayesian Multi-label Learning with Sparse Features and Labels, and Label Co-occurrences” co-authored with He Zhao, Piyush Rai, and Wray Buntine is published in AISTATS 2018

Grants

Education Background

  • Post-doc Research Fellow, Department of Computing, Macqaurie University, 2012 - 2015, working with Prof. Mark Johnson
  • Ph.D. Computer Science, The Australian National University, Australia, 2008 - 2011
    Supervisors: Prof. Wray Butine, Dr. Huidong Jin
  • B.S. Information Technology with 1st class honours, The Australian National University, Australia, 2007
    Supervisor: Dr. Huidong Jin
  • B.S. Information and Communication Technology, Flinders University, Australia, 2006