Projects
-
Linux Technology Centre, IBM,Bangalore (Jan '05 to Jun '05)
Project: Analysis of Reference Counting Algorithms for RCU based Lock-free Data Structures in Linux Kernel
Abstract : With the advent of shared-memory parallel (SMP) systems, an additional level of complexity was added to locking primitives. Read-copy update(RCU) like many other synchronization primitives was hence developed as a reader-writer synchronization mechanism where readers can incur zero synchronization overhead. RCU provides improved performance in the case of read-mostly data structures. RCU is implemented in the linux kernel(eg. dcache, route cache) in different forms. However, most of them use reference counting for keeping track of the references to the shared data structure. Some of these RCU algorithms and the different atomic synchronization operations that they use are analyzed during the course of this work. The lookup performance of these algorithms is measured under different scenarios and is compared with respect to each other.
For a complete report, please click here.
-
Technology Incubation Centre, IBM,Bangalore (July '04 to Dec '04)
Project: Planning Algorithms and Knowledge Representation for Autonomic Computing
The aim of this project was to design a knowledge base for Autonomic Computing and design and implement a multiagent planning algorithm. Knowledge representation standards like RDF, DAML, OWL and ontologies like CIM and Gene Ontology were explored. Analysis of Ontology API's and editors like IBM Ontology Management Toolkit (Snobase), Protege, Jena and Chimera was done for creating and querying ontologies. Analysis of different querying languages like DQL and RDQL was also done. OWL was selected for building the ontology since it captured all the semantics of the domain. An in depth analysis of planning algorithms like STRIPS, HTN Planning Algorithms and others was done in order to design a planning algorithm for the problem of Autonomic Computing. Finally, a planning engine was designed and implemented using Java on Websphere platform.
For a complete report, please click here.
Indian Statistical Institute, Calcutta (May '04 to July '04)
Project : Analysis of Markov models for clustering DNA sequence data
By using Statistical techniques on sequence data, it was established that though the DNA sequence data do not quite follow Markov models, transition probability or word probability parameters can be used to successfully cluster the sequences (or organisms) perhaps because certain transition probabilities are organism specific. Statistical concepts like correlation coefficients, stationary distributions, estimation techniques and others were understood and then used in establishing the results.
For a complete report, please click here.
-
Microsoft Research, Redmond (June '03 to Aug '03)
Project : Bayesian Document Classification
A document classification\clustering engine based on the text contents of the documents was developed. The Naïve Bayes Classifier and Similarity trees were used in clustering/classification. The extraction of the text from documents was done using TfIdf factor for the words. Phrase extraction from the documents was done using mutual information between words. Finally, a neat user interface was developed for the application.
-
Bhabha Atomic Research Centre, Bombay (May '02 to July '02)
Project : Characterization of 3D objects Constructed from Sliced Image Files
Cancer cells were the 3dobjects that were reconstructed and characterized. Sliced images of cancer cells were taken as input. Reconstruction of cancer cells was done using Ray tracing and Marching Cubes algorithms. Characterization of cancer cells was performed based on the important property of Cancer cells that they have more number of protrusions or bumps on their surface as compared to normal cells. Hence, by analysis of the surface roughness of cancer cells, it was determined if the given cell was cancerous or not.