PhD Student
Computer Science and Engineering
University of California, Riverside

Home | Research | Links | Courses | Contact

 

my pic

Home

Research

Links

Courses

Contact
Hi, My name is Abdullah Al Mueen. I am a PhD candidate at the department of Computer Science and Engineering in University of California, Riverside. Here is my resume.

Dr. Eamonn Keogh is my adviser. I have mastered techniques on finding time series motifs exactly from in-memory, disk resident and online data. I am currently exploring online algorithms for time series data mining.

In summer 2009, I was a research intern at Microsoft Research, Redmond and worked with Dr. Suman Nath.

In spring 2010, I was a research intern at Microsoft Research, Redmond with Dr. Naga Govindaraju.

In summer 2011, I was a research intern at HP Labs, Palo Alto, CA  in the Information Analytics Lab.

Previously, I got my bachelor degree in computer science and engineering from Bangladesh University of Engineering and Technology (BUET).

Research
I am broadly interested in Data Mining and Pattern Recognition.

I like playing with large data and their inherent structures. I am particularly interested in high dimensional data objects like time series, images, XML documents, videos and their
  • Similarity Measures
  • Motifs/Patterns inside the Data
  • Geometric Properties, Lower Bounds
  • Joins, Nearest Neighbor, Correlation Search
My Selected Publications are
  • Mining Massive Archive of Mice Sounds with Symbolized Representations, Jesin Zakaria, Sarah Rotschafer, Abdullah Mueen, Khaleel Razak, Eamonn Keogh to appear in the Proceedings of SDM 2012. [pdf]

  • Image Mining of Historical Manuscripts to Establish Provenance, Bing Hu, Thanawin Rakthanmanon, BIlson Campana, Abdullah Mueen, Eamonn Keogh, to appear in the Proceedings of SDM 2012. [pdf]
  • Logical-Shapelets: An Expressive Primitive for Time Series Classification, Abdullah Mueen, Eamonn Keogh, Neal Young, In the Proceedings of ACM SIGKDD 2011. pp. 1154-1162.  [pdf]  

    Here is a supporting webpage for this paper that contains all the datasets.

  • Accelerating Dynamic Time Warping Subsequence Search with GPUs and FPGAs, Doruk Sart, Abdullah Mueen, Walid Najjar, Vit Niennattrakul, Eamonn Keogh, In the Proceedings of IEEE ICDM 2010. pp. 1001-1006. [pdf][slide]

    I have been awarded the NSF     student travel grant to attend the conference in Sydney, Australia.
     
  • Online Discovery and Maintenance of Time Series Motif, Abdullah Mueen, Eamonn Keogh, In the Proceedings of ACM SIGKDD 2010. pp. 1089-1098. [pdf][slide]

    Here is the supporting webpage for this paper. This youtube demo shows the algorithm in action. I have been awarded the student travel grant for this conference.
  • Fast Approximate Correlation for Massive Time-Series Data, Abdullah Mueen, Suman Nath, Jie Liu, In the Proceedings of ACM SIGMOD 2010. pp. 171-182. [pdf][slide]

    I have been awarded the student travel grant to attend this conference.

  • Finding Time Series Motifs in Disk-Resident Data, Abdullah Mueen, Eamonn Keogh, Nima Bigdely-Shamlo, In the Proceedings of IEEE International Conference on Data Mining, pp. 367-376, ICDM 2009. [pdf][slide]

    Here     is the supporting webpage for this paper. An expanded version of this paper has been accepted in Data Mining and Knowledge Discovery Journal (DMKD 2010). This youtube video shows an example of mlutidimensional motif.

  • Exact Discovery of Time Series Motifs, Abdullah Mueen, Eamonn Keogh, Qiang Zhu, Sydney Cash, Brandon Westover, In the Proceedings of SIAM International Conference on Data Mining, pp. 473-484, SDM 2009. [pdf]

    Here is a supporting webpage for this paper that contains all the datasets.

  • A Heuristic Algorithm for Individual Haplotyping with Minimum Error Correction, Abdullah Al Mueen, Md. Shamsuzzoha Bayzid, Md. Maksudul Alam, Md. Saidur Rahman, In the Proceedings of International Conference on BioMedical Engineering and Informatics, pp. 792-796, BMEI 2008.
The complete list is DBLP list.

My Research Statements (will be growing in the coming years)
  • Parameters are killers of algorithms. Only one unintuitive parameter is enough to kill an algorithm.
  • "Simple techniques have the amazing power of generality" -- My Adviser. I beleive it strongly.
  • Forget about time complexity, if an algorithm can discover meaningful/interpretable results.
  • Clever techniques for single CPU may be worthless while converting to run in a parallel environment.
  • Data Mining and large scale data analysis are synonymous to me. So small scale data analysis is not data mining.
  • Showing a new problem is difficult/impossible to solve is better than showing an old problem has another better/improved solution.
  • Adding more and more relevant dimensions to the search space with good pruning strategy is key for building more intelligent systems.
  • Special cases need special attention.
  • Avoid curse of knowledge by switching domains.

My research is being funded by the UCR Computational Anthropology Project. (NSF 0803410)

Links

Graduate Courses
Not available...

Contact

or

Room 368,
Engineering Building Unit 2,
Univerity of California, Riverside.
California 92521.