Maunendra Sankar Desarkar
PhD Student
Department of Computer Science and Engineering
Indian Institute of Technology Kharagpur
Kharagpur, West Bengal - 721302, India
Email: maunendra @ cse.iitkgp.ernet.in
[ Research |
Publications |
Activities |
Projects |
Internships |
Work Experience |
Education |
Awards ]
Research
Information Retrieval, Data Mining, Machine Learning
Supervisor: Prof. Sudeshna Sarkar
Publications
-
Maunendra Sankar Desarkar, Sudeshna Sarkar. ``Rating Prediction Using Preference Relations Based Matrix Factorization ''
FactMod 2012 Workshop on Matrix Factorization Techniques for User Preference and Skills Modeling at UMAP 2012,
Montreal, July 16-20, 2012.
- Maunendra Sankar Desarkar, Sudeshna Sarkar.
``User Based Collaborative Filtering with Temporal Information for Purchase Data'' in
4th International Conference on Knowledge Discovery and Information Retrieval, KDIR 2012,
Barcelona, Oct 4-7, 2012.
- Abir De, Maunendra Sankar Desarkar, Niloy Ganguly and Pabitra Mitra.
``Local Learning of Item Dissimilarity Using Content and Link Structure '' in
6th ACM Conference on Recommender Systems, RecSys 2012,
Dublin, September 9-13, 2012.
- Maunendra Sankar Desarkar, Sudeshna Sarkar, and Roopam Saxena.
``Preference Relation Based Matrix Factorization for Recommender Systems'' in
20th International Conference on User Modeling, Adaptation and Personalization, UMAP 2012,
Montreal, July 16-20, 2012.
- Maunendra Sankar Desarkar, Rahul Joshi, and Sudeshna Sarkar.
``Displacement Based Unsupervised Metric for Evaluating Rank Aggregation'' in
4th International Conference on Pattern Recognition and Machine Intelligence,
PReMI 2011, Moscow, June 25 - July 1, 2011.
- Maunendra Sankar Desarkar, Sudeshna Sarkar, and Pabitra Mitra.
``Aggregating Preference Graphs for Collaborative Rating Prediction'' in
4th ACM Conference on Recommender Systems, RecSys 2010, Barcelona, Sep
26-30, 2010.
Academic Activities
- Reviewer: IEEE TechSym 2011
- Teaching Assistant for Machine Learning (Spring 2010)
- Teaching Assistant for Discrete Structures (Autumn 2010)
- Teaching Assistant for Machine Learning (Spring 2009)
- Teaching Assistant for Information Retrieval (Autumn 2009)
- Teaching Assistant for Programming and Data Structures (Spring 2008)
- Teaching Assistant for Algorithms-II (Autumn 2008)
- Teaching Assistant for Introduction to Data Structures and Algorithms (Autumn 2004 )
- Teaching Assistant for ESC101 - Introduction to computing (Spring 2004)
Projects
- On Application of Multi-parameter Auctions to Page Allocation in Distributed Shared Memory Multiprocessors
- In multiprocessor systems, agents (processors) compete with each
other for different resources (main memory, cache etc).
Algorithms which take allocation decisions are generally complex in
nature. We investigated how the system would perform if
the agents are allowed to behave on their own, in the presence of any
central authority (following game-theoretic approaches)
and how to design simple allocation algorithms (with performance
guarantees) that are easy to implement in hardware. We focused
on the page allocation problem and identified a mechanism design based
approach to solve the problem. We modeled the problem as
a multi-parameter auction where the processors are the agents and the
pages are the goods to be allocated. The page allocation
algorithm serves as the central authority. For such an auction-based
setting, we established the necessary and sufficient
conditions for ensuring the truthfulness of the agents. Based on this,
we developed page allocation algorithms for DSM multiprocessors
and theoretically analyzed the performance of the algorithms.
- M. Tech. Thesis
- Spervisor: Prof. Mainak Chaudhuri
- Data Profiling for ETL Applications
- Data cleaning is an important phase of any ETL process. This prevents erroneous data from entering the
target database. In data profiling we try to find some rules which are followed by the correct data and use
that knowledge to filter out dirty data from the database. In this project we built a tool that is able to
extract some simple rules from a clean database. We also developed an operator library containing some
operators those would be available for Europa and be helpful in designing ETL processes on Europa platform.
- During Summer Internship at IBM India Research Labs in 2005
- Mentor: Natwar Modani
- Developing an optimal algorithm for Two Dimensional String Matching in Mesh Connected Parallel Network
- In this work, we came up with a simple and efficient algorithm for
two dimensional string matching on a mesh
connected network of processors. The algorithm first preprocesses the
pattern (of size m X m) in
O(m^2 log |\Sigma|) time where \Sigma is the alphabet. It then finds all
and exact occurrences of the pattern in O( \frac{n^2}{p^2})
time (which is optimal) where p^2 processors are used and dimension of
the text is n X n, provided m < \frac{n}{p}.
- In Parallel Semi-Numerical and Non-Numerical Algorithms course in Spring 2004
- Instructor: Prof. Phalguni Gupta
- Segmentation of Multi Spectral Images using Active SVM
- In this work we took brain images taken in different frequency bands and used active learning to find different segments.
For classification we used SVM.
- In Knowledge Discovery course in Autumn 2004
- Instructor: Prof. Pabitra Mitra
Internships
- Data cleaning for ETL Applications at IBM India Research Lab from May 2005 to July 2005
Work Experience
-
Sybase India Pvt. Ltd.
- Duration: July 2006 to July 2008
- Role: Worked as software engineer in Sybase IQ DDL team.
- Projects worked on: Data loading and index population, message log management, table partitioning and tablespaces.
Education
- Pursuing PhD in Department of Computer Science and Engineering, Indian Institute of Technology Kharagpur
- Thesis topic: aggregating structured opinions for search and recommendation
- Supervisor: Prof. Sudeshna Sarkar
- M. Tech. in Computer Science and Engineering from Indian Institute of Technology Kanpur (2004-2006)
- CGPA: 8.84
- Thesis: On Application of Multi-parameter Auctions to Page Allocation in Distributed Shared Memory Multiprocessors
- Supervisor: Prof. Mainak Chaudhuri
- B. E. in Computer Science and Engineering, University of Burdwan (2000-2004).
- Percentage: 88.6%
- Project: Delay Analysis of a Heterogeneous Multi-server Network
- Higher Secondary Examination under West Bengal Council of Higher Secondary Education (2000)
- Madhyamik Examination from West Bengal Board of Secondary Education(1998)
Awards
- Honorable Mention Award in Yahoo! Key Scientific Challenge 2012.
- Recipient of Microsoft Research India PhD Fellowship for the period 2009-2013.
- Secured 99.73 percentile (AIR 95) in GATE 2004.
- Top ranker of The University of Burdwan in BE (including all branches - Winner of University Gold Medal).
- Ranked 26th in state in Higher Secondary Examination.
- Ranked 45th in state in Madhyamik Examination.
- Member of Organizing committee of Horizon 2003 - a state level
technical festival organized by Dr. B. C. Roy Engineering college.
- 2nd in Programming contest in Arohan 2004 - a national level programming contest organized by NIT Durgapur.
- Finalist of Programming contest in Arohan 2003 - a national level programming contest organized by NIT Durgapur.
Last updated: June 2012