Embed Size (px)
Citation preview
Akhil Arora
ContactInformation
Researcher, Text & Graph Analytics Group T +91-9980378574 (India)Xerox Research Centre India B [email protected], [email protected], India WWW : http://cse.iitk.ac.in/∼aarora
ResearchInterests
Large Scale Data Mining: Graph Mining, Social Network Analysis; Databases: Indexing & QueryingLarge Graphs, Text and High Dimensional Databases
Education Masters in Computer Science CGPA : 8.67/10.0 (Rank: 3/39)Indian Institute of Technology (IIT), Kanpur, India July 2011 – June 2013Advisor: Prof. Arnab Bhattacharya
SelectedPublications
• VLDB 2016: Satyajit Bhadange, Akhil Arora∗, Arnab Bhattacharya: GARUDA: A Systemfor Large-Scale Mining of Statistically Significant Connected Subgraphs, In: Proc. of Interna-tional Conference on Very Large Databases, 2016 (Demonstrations Track)
• SIGMOD 2016: Akhil Arora†, Sainyam Galhotra†, Shourya Roy: Holistic Influence Maxi-mization: Combining Scalability and Efficiency with Opinion-Aware Models, In: Proc. of ACMInternational Conference on Management of Data, 2016
• WWW 2015: Sainyam Galhotra†, Akhil Arora†, Srinivas Virinchi, Shourya Roy: ASIM: AScalable Algorithm for Influence Maximization under the Independent Cascade Model, In: Proc.of ACM International Conference on World Wide Web, 2015 (Poster: Companion Volume)
• SIGMOD 2014: Akhil Arora, Mayank Sachan, Arnab Bhattacharya: Mining StatisticalltySignificant Connected Subgraphs in Vertex Labeled Graphs, In: Proc. of ACM InternationalConference on Management of Data, 2014
ProfessionalActivities
• Co-Chair – Network Data Analytics (NDA’16) Workshop (Co-located with SIGMOD 2016);Research Programming Challenge at COMAD 2014.
• Program Committee Member – GDAM 2015, XRCI Open 2015.• Reviewer – VLDB 2016, KDD (2016, 2015), SDM (2016, 2015), CIKM (2015, 2014), COMAD
(2016, 2014, 2013), CoDS 2015, Journal of Information Science 2014.• Organizing Committee – XRCI Open (2015, 2016).• Co-founder and organizer of Special Interest Group in Data (SIGDATA) at IIT Kanpur,
which meets weekly to discuss relevant advances in Databases and Data Mining
Work Experience Researcher, Xerox Research Centre India (XRCI)Member of Text and Graph Analytics Group August 2014 – PresentLeading projects on devising scalable algorithms to solve a gamut of complex real-world problems inthe area of databases, data mining and machine learning. Work done here led to several publicationsin SIGMOD, VLDB and WWW.
Software Engineer, Intel Corporation, Bangalore, IndiaMember of Security and Vulnerability Hacking Group July 2013 – July 2014Worked on research problems in security while performing white hat hacking on internal Intel prod-ucts, security code reviews, assessments, and code assisted penetration. Developed a frameworkwhich was published in Black Hat 2014.
∗Corresponding Author†Equal Contribution
Honors andAwards
• Invited to present our SIGMOD 2014 paper in the premier papers track at the 20th Interna-tional Conference on Management of Data (COMAD 2014).
• Won the First Prize in the Adobe Data Mining Competition at IIT, Madras.• Won the Fifth Prize in Scalable String Similarity Search/Join workshop: EDBT, 2013.• Awarded the Overall Best Hack prize in Yahoo! HackU!, 2012 at IIT, Kanpur.• Won the Second Prize in the 10th ImageCLEF: Plant Identification Task, 2012.• Among 7 researchers worldwide to get ELIAS Sponsorship for attending CLEF, 2012.• Best Project Award for HiPhi: Approximate kNN Search in High-Dimensional Spaces.• Best Project Award for FriendMiner: Inferring Relationship Based on Mobile Phone Data.• Ranked Third in the department, among the entrant batch of M.Tech and Phd programme.• All India Rank(AIR) 369, GATE 2011 (percentile 99.73), category: Computer Science.• State level ‘Science Award’ for securing highest marks in Mathematics in Secondary Ex-
amination (ICSE), 2004.• Uttar Pradesh State Merit Scholarship for Academic Excellence in Secondary Examination
(ICSE), 2004.
Invited Talks • Holistic Influence Maximization: Scalability and Efficiency with Opinion-Aware Models◦ Indian Institute of Technology (IIT), Delhi, India September, 2016◦ The Northcap University (NCU), Gurgaon, India September, 2016◦ University of California, Santa Barbara, USA June, 2016◦ Facebook Inc., Menlo Park, USA June, 2016◦ Palo Alto Research Centre (PARC), USA June, 2016◦ Indian Institute of Technology (IIT), Kanpur, India March, 2015
• ASIM: A Scalable Algorithm for Influence Maximization under the Independent Cascade Model◦ Indian Institute of Technology (IIT), Kanpur, India February, 2015
• Mining Statisticallty Significant Connected Subgraphs in Vertex Labeled Graphs◦ Palo Alto Research Centre (PARC), USA September, 2015◦ 20th Int. Conf. on Management of Data (COMAD 2014) December, 2014◦ Xerox Research Centre Europe, Grenoble, France October, 2014
Masters Thesis Mining Statistically Significant Subgraphs Spring 2012 – Spring 2013Worked with Prof. Arnab Bhattacharya to develop the first ever scalable algorithm to mine statisti-cally significant subgraphs from a single large graph, with a wide-variety of applications ranging fromCommunity Detection to Mining Spatial Colocations, Hotspot Detection and many more. Insteadof the more commonly used frequency, we use the more involved p-value/chi-square statistic as anobjective function to mine interesting patterns that deviate significantly from the expected. Thiswork was published as a paper in the research track at SIGMOD 2014.
OtherPublications,PatentApplications andDisclosures
• Akhil Arora, Manoj Gupta, Neeta Pande, Sainyam Galhotra, Shourya Roy: System for Iden-tifying Root Causes of Churn for Churn Prediction Refinement, Filed: USPTO (2016)
• Sainyam Galhotra†, Akhil Arora†, Shourya Roy: Holistic Influence Maximization: CombiningScalability and Efficiency with Opinion-Aware Models, In: NEDB North East Database Day(Poster), 2016
• Akhil Arora, Manoj Gupta, Shourya Roy: Transforming a Knowledge Base into a MachineReadable Format for an Automated System, Filed: USPTO (2015)
• Akhil Arora, Sainyam Galhotra, Srinivas Virinchi, Shourya Roy: Methods and Systems forIdentifying Target Users of Content, Filed: USPTO (2015)
• Deepali Semwal, Sonal Patil, Sainyam Galhotra, Akhil Arora, Narayanan Unny: STAR: Real-time Spatio-Temporal Analysis and Prediction of Traffic Insights using Social Media, In: ACMIKDD Conference on Data Sciences (CoDS), 2015
• Akhil Arora, Sumanth Naropanth: Android Kernel and OS Security Assessment with IronCrow, Black Hat Europe, 2014
• Shashwat Mishra, Tejas Gandhi, Akhil Arora, Arnab Bhattacharya: Efficient Edit Distancebased String Similarity Search using Deletion Neighborhoods, In: Proceedings of the Joint EDBT
/ICDT 2013 Workshops• Akhil Arora, Ankit Gupta, Nitesh Bagmar, Shashwat Mishra, Arnab Bhattacharya: A Plant
Identification System using Shape and Morphological Features on Segmented Leaflets: Team IITK,CLEF 2012 In: CLEF 2012 (Online Working Notes/ Labs/Workshop)
ResearchProjects
GaBiD: Graphical Analysis of Big data Xerox ResearchProject Co-lead with Dr. Shourya Roy Jan 2015 – Present• Responsible for design and development of a churn analytics suite consisting of novel graph based
algorithms for (1) prediction, (2) root-cause identification and (3) prevention; to holistically solvethe problem of Customer Churn.
KEO: Knowledge Extraction and Organization Xerox ResearchProject Co-lead with Dr. Sumit Negi and Dr. Shourya Roy Jul 2014 – Dec 2015• Responsible for design and development of automated methods for knowledge base construction
by performing information extraction and organization from enterprise content.
STaCHIT: Smart TimeLine and Chit-Chat (Overall Best Hack) Yahoo! HackUProf. Arnab Bhattacharya, IIT Kanpur; Dr. Muthusamy Chelliah, Yahoo Research Fall, 2012• Developed a system for smart rendering of web-content, comprising of a rich user interface with
loads of novel features like timelined news reader, storification and smart chat assist.
ImageCLEF: Plant Identification System (Winner: ImageCLEF) Machine LearningProf. Arnab Bhattacharya, IIT Kanpur (Independent project in summer) Summer, 2012• Developed a framework for automatic categorization of plant types using shape, morphological
and novel tooth features with a Random Forest classifier, beating the state-of-the-art in quality.
HiPHi: Indexing High Dimenisonal Spaces for Approximate kNN Search DatabasesProf. Arnab Bhattacharya, IIT Kanpur Spring, 2012• Proposed a novel data-structure, RDB-tree, that modifies B+-tree using a combination of space-
filling curves and pivots; providing orders of magnitude improvement over the state-of-the-art.
FriendMiner: Relationship Inference using Mobile Phone Data Data MiningProf. Arnab Bhattacharya, IIT Kanpur (Independent Research) Fall, 2011 – Spring, 2012• Proposed a novel strategy using a combination of proximity information from Bluetooth logs and
location features, with ensemble methods for relationship classification.
On the Fusion of Periocular and Iris Biometrics Biometric SystemsProf. Phalguni Gupta, IIT Kanpur Fall, 2011• Novel strategy to fuse iris and periocular biometrics resulting in improvements in the recognition
rate in case of non-ideal imagery.
Teaching andMentoringExperience
• Teaching Assistant, Introduction to Computing, IIT Kanpur Fall 2011, Spring 2012• Teaching Assistant, Data Mining, IIT Kanpur Fall 2012• Teaching Assistant, Database Management Systems, IIT Kanpur Spring 2013
• Masters Thesis Co-Supervisor, IIT Kanpurco-supervised students and masters thesis as follows– Piyush Kumar, Indexing High-Dimensional Databases Spring 2015 – Spring 2016– Sakshi Sinha, Indexing High-Dimensional Databases Spring 2015 – Spring 2016– Satyajit Bhadange, Distributed Significant Subgraph Mining Spring 2015 – Fall 2015
• Mentor, IIT Madrasmentored PhD thesis as follows– Jithin Valchery, PhD student (IIT Madras) Aug. 2015 – Present
working on a novel problem of contextual graph querying.
• Mentor, Xerox Research Centre India (XRCI)mentored intern projects as follows– Prajna Upadhyay, PhD student (IIT Delhi) Aug. 2015 – Dec. 2015
designed a system for automatic knowledge extraction and organization from enterprise data.– Srinivas Virinchi, PhD student (UMD, College Park) Aug. 2014 – Dec. 2014
developed a scalable algorithm for influence maximization in large networks.
Technical Skills Programming languages: C, C++, Python, SQL, Java, R, PrologPackages used: Numpy, Scipy, BoostApplications and Tools: Weka, LATEX, OpenCV, Matlab, GNU Octave, Shell scripting, GNUPlot, OpenOffice, MS Office, Verilog(HDL)Worked with large graph datasets: ranging from a few thousand nodes and edges (NetHEPT– 15K nodes and 62K edges), to a million nodes and billion edges (Twitter – 41M nodes and 1.5Bedges; Friendster – 65M nodes and 1.8B edges)Operating systems: Linux, WindowsWeb Designing: HTML, XHTML, XML, PHP, CSS, Javascript
References Prof. Arnab BhattacharyaAssociate ProfessorDepartment of Computer ScienceIIT Kanpur, IndiaB [email protected] [email protected]
Prof. Sayan RanuAssistant ProfessorDepartment of Computer ScienceIIT Madras, IndiaB [email protected]
Dr. Shourya RoySenior Research ScientistXerox Research Centre IndiaBangalore, IndiaB [email protected]
Dr. Manish GuptaDirector and Vice PresidentXerox Research Centre IndiaBangalore, IndiaB [email protected]
