Introduce MIKETM lab

1
Introduce MIKETM lab

• Massive Information Knowledge Engineering
• Department of Computer Engineering
• Kasetsart University, Bangkok
• Thailand.

http//mike.cpe.ku.ac.th
GoogleTM not approved yet, but soon!
V2006
2
Current members
3
Something deserved to remember

• Founded on 1998 in Dr.Arnons office, 16m2 area.
• Dr.Putchong named us, MIKE.
• Got another 20m2 from old PRG area of Dr.Putchong
  in 1999.
• 3 master degree students, 8 Eng. students.
• The first generation of computing equipments was
  collected from waste area of the other labs in
  the department.
• First fund, 12000 baht, from Dr. Sumnuk, the head
  department to buy a hard disk.

4
Then and now

• MIKETM area on the 7th floor, department of
  computer engineering, around 50m2.
• 3 Ph.D. students, but 6-7 Ph.D. students during
  the meeting.
• 1 under the royal golden jubilee program, 1 from
  CRN, and 1 from Thai gov. fund.
• Others come from inter-lab (HPCNC, KU-CS) and
  inter-university (Chula) collaboration.
• 4 master students (2 active and 2 passive) and
  around another 10 golden fishes if we counts them
  as always. _)
• No Eng. Students for a moment.

5
What we do

• Anything concerning search and management of both
  information and knowledge on the Internet.
• Gathering and storage
• Management (analyze, synthesize, index, )
• Utilization (Search Engine interface)
• Parallel and cluster computing of the above
  applications.
• Distributed and grid computing of the above
  applications.
• Bio-informatics.

6
Information Gathering Storage

• Topic specific web spider (Journal link ?)
• Collect only relevant pages of interested topics
  from the Internet.
• Use machine learning technique to learn from
  passed crawling experiences and build a new
  knowledge.
• Language specific web spider
• TGIST 2006 fund under the collaboration with
  rdi5-NECTEC. ?
• Dig the Internet to find web page in a specific
  language.
• Region web spider
• Cooperative information gathering using Grid
  computing infrastructure.

7
Information Management

• Web link analysis
• Web link spammer detection and correction
• Identify spamming cluster and recalculate un-bias
  scores
• Parallel adaptive PageRank computation
• Mixed MPI Threads model on dual Intel
  quadruple cores
• Age-based PageRank computation
• Calculate PR according to the pages age
• Web crawling refresh model
• Model the update behavior of web pages on the
  Internet
• Content aware PageRank computation

8
Information Management (2)

• Indexing
• PWORMs (Vector Space based retrieval system)
• Lucene with PageRank
• Image retrieval
• Automatic images annotation
• Web mining (under inter-lab and inter-U
  collaboration)
• Parallel data mining
• Distributed data mining algorithm

9
Information Utilization
10
Bio-Informatics

• Classification on Medical dataset
• Predict trend of prognosis
• Discovering new knowledge
• Extract new knowledge of medical term for
  treatments from medical database.

11
Computing resources that we can dig

• From HPCNC
• 6 nodes of dual Intel quadruple cores, 3.2 GHz,
  64 Gbtye.
• From the KU computing service dept.
• 32 nodes of dual Opteron 240 cores, 1.4Ghz, 323
  Gbyte.
• ApGrid F32 cluster
• 260 nodes of dual Xeon cores, 3GHz, 2603 Gbyte.

12
Publications

• Ask DBLP
• Ask citeseer
• Ask Google
• Ask Vivisimo
• or ask www.kartoo.com if you like.

13
Contact us

• Massive Information Knowledge Engineering
• Dept. of computer engineering
• Faculty of Engineering
• Kasetsart University, Bangkok, Thailand.
• arnon_at_mikelab.net
• http//mike.cpe.ku.ac.th
• ????????????????????????? ?????????????????????
• ???? E707 ???? 7 ???????????????????????????????
  
• ????????????????? ?????????????????????? ???????
  10900
• ???????? 02-9428555 ??? 1429, 1449

Thank you for your attention.