|Address:|| 2707 Portland Street, Apartment 208,|
Los Angeles, CA-90007
I am a graduate student at University of Southern California and I'm currently pursuing second semester of my master's degree here.
I am currently looking for full time opportunities for 2015.
I am interested in the areas of Artificial Intelligence, Algorithm Design, Natural Language Processing and Machine Learning.
I participated in Google Summer of Code 2014 with National Resource for Network Biology and implemented bi-clusters and their vizualization for their network vizualization framework- Cytoscape.
Developed a prototype for a secure on cloud encryption and storage service with on premise key management. Implemented the service on OpenStack using Swift and Keystone APIs.
Implemented bi-clustering algorithms in Cytoscape, a platform for visualization of complex networks
Implemented Cheng and Church, BiMine, BicFinder and optimal visualization of bi-clusters using heatmap re-ordering.
Implemented fuzzy clustering algorithms in Cytoscape, a platform for visualization of complex networks.
Implemented the visualization of fuzzy clusters using a combination of membership edges and degree of membership.
Created a framework for extracting user information (preferences, dislikes, health related information etc.) for the user models in conversational agents, aimed at assisting the elderly and isolated in need of special care.
Worked on information annotation and extraction using dependency relations in conversation transcripts using UIMA framework and ClearTK (framework for developing statistical NLP components).
Created an online question bank organizer, for generating various question papers of comparable difficulty level. Generated question papers utilizing an SQL database in a randomized fashion, making necessary changes in the database
Developed a framework for detection and analysis of events in a match using tweets. Clustered the tweets corresponding to each team, performed sentiment classification and identified the key players. Segmented tweets into events based on the burst rates and created a summarizer by ranking tweets in a segment.
Implemented an averaged perceptron classifier and used it to develop a POS tagger and an NER module. Developed a language model based framework for homophone error detection and correction. Implemented a Naïve Bayes text classifier for spam filtering and sentiment analysis.
Designed a Spatial Search Engine for searching queries in geo-tagged documents, indexed with Apache Solr. Configured Apache Nutch to crawl the FBI vault's mirror website and extracted PDFs, utilizing Apache Tika. Used Apache Tika for extracting the number of hits corresponding to a set of queries from a vault of documents. Further enhanced the mechanism by incorporating Levenshtein distance in query search.
Project involves designing and implementing an intelligent robot, capable of navigating a path in a crowded environment.
Implementing a mechanism to make the robot learn from its own experience, to find a target efficiently.
Created an android app for the purpose of weather search.
The app uses the weather feed returned by yahoo's weather api which is converted into xml format by a php script running on
Amazon's cloud server. A java servlet generates JSON string from the xml data and this is utilized by the app.
The app also provides functionality to post and share the weather information to user's facebook profile.
Implemented a new algorithm for clustering web documents using Latent Semantic Analysis and Minimum Cuts in graphs.
Implemented a ranking algorithm which ranks web documents with respect to a submitted query.
Created an applet for optimizing the downstream synthesis of human insulin in terms of the revenue and time.
The algorithm incorporated the bounds on the various parameters in the multiple stages and their interdependencies.
Primary author of the chapter Affective Behavioral Cognitive Learner Modeling for Army Research Laboratory's book- Design Recommendations for Intelligent Tutoring Systems, Volume 1: Learner Modeling.
The chapter presents a framework for modeling users of e-learning systems that integrates inductive and abductive reasoning over observations including the learner's past and current behavior to develop a joint model for predicting emotions, behaviors, and cognitive states. This ABC Learner Model follows an approach that learners' behavioral responses can be a path to predict, recognize, and interpret their affective state. These behavioral responses are analyzed using a cognitive theory of emotions, which gives us inferences about the possible affective states of the learner. An appraisal component of the model relies on the desirability of events given the learner's objectives, the resulting affective and cognitive states predicted to result from the events, and the consequent behaviors expected.
Get the Chapter here
Address: 2707 Portland Street, Apt 208,Los Angeles, CA-90007 Phone: 347-824-8524 email: firstname.lastname@example.org