My background is in computer science and computational neuroscience with a focus on machine learning, signal processing and applied math modeling. My research interests are varied and the fields I've worked on include Machine Learning and Data Mining, Signal Processing, Psychophysics, Cryptography (Secure Multiparty Computation), Wireless and Network Algorithms, Graph Theory and other areas of Computer Science.
EDUCATION
2003 - 2007
Boston University, Boston, MA - 3.96/4.00
PhD, Cognitive and Neural Systems (CNS)
Thesis Title: Latent Variable Framework for Modeling and Separating Single Channel Acoustic Sources.
Advisor: Prof. Barbara Shinn-Cunningham
1999 - 2003
Birla Institute of Technology and Science, Pilani, India - 9.39/10.00
BE (Honors) in Computer Science
PUBLICATIONS
Journal Papers
Missing Data Imputation for Time-Frequency Representations of Audio Signals. P Smaragdis, B Raj, MVS Shashanka.
Journal of Signal Processing Systems, To Appear.
Probabilistic Latent Variable Models as Non-Negative Factorizations. MVS Shashanka, B Raj, P Smaragdis.
Computational Intelligence and Neuroscience, Special Issue on Advances in Non-Negative Matrix and Tensor Factorizations, May 2008.
A Framework for Secure Speech Recognition. P Smaragdis, MVS Shashanka.
IEEE Transactions on Audio, Speech and Language Processing, Vol 15, No. 4, May 2007, p1404-1413.
Optimal Skewed Data
Allocation on Multiple Channels with Flat Broadcast per Channel.* E Ardizzoni, AA Bertossi, MC Pinotti,
S Ramaprasad, R Rizzi, MVS Shashanka.
IEEE Transaction on Computers, Vol 54, No. 5, May 2005, p558-572. *Authors listed in alphabetical order.
A Characterisation of Optimal Channel Assignments for Wireless
Networks Modelled as Cellular and Square Grids. MVS Shashanka, A Pati, AM Shende.
Mobile Networks and Applications, Vol 10, Issue 1-2, Feb-Apr 2005, p89-98.
Conference Proceedings
Simplex Decompositions using SVD and PLSA. MVS Shashanka, M Giering.
Intl Conf on Pattern Recognition Applications and Methods, Vilamoura, Portugal, Feb 2012.
Copula Functions for Learning Multimodal Densities with Non-linear Dependencies. A Tewari, MVS Shashanka, MJ Giering.
NIPS Workshop on Copulas in Machine Learning, Sierra Nevada, Spain, Dec 2011.
A Fast Algorithm for Discrete HMM Training using Observed Transitions. MVS Shashanka.
IEEE Intl Conf on Acoustics, Speech and Signal Processing, Prague, Czech Republic, May 2011.
A Privacy Preserving Framework for Gaussian Mixture Models. MVS Shashanka.
IEEE Intl Workshop on Privacy Aspects of Data Mining, Sydney, Australia, Dec 2010.
Probabilistic Latent Component Analysis for Gearbox Vibration Source Separation. J Isom, MVS Shashanka, A Tewari, A Lazarevic.
Annual Conference of the PHM Society, Portland, Oregon, Oct 2010.
Topic Models for Audio Mixture Analysis. P Smaragdis, MVS Shashanka, B Raj.
NIPS Workshop on Applications for Topic Models: Text and Beyond, Vancouver, Canada, Dec 2009.
A Sparse Non-parametric Approach for Single Channel Separation of Known Sounds. P Smaragdis, MVS Shashanka, B Raj.
Neural Information Processing Systems Conference (NIPS), Vancouver, Canada, Dec 2009.
Simplex Decompositions for Real-Valued Datasets. MVS Shashanka.
IEEE Intl Workshop on Machine Learning and Signal Processing, Grenoble, France, Sep 2009.
Missing Data Imputation for Spectral Audio Signals. P Smaragdis, B Raj, MVS Shashanka.
IEEE Intl Workshop on Machine Learning and Signal Processing, Grenoble, France, Sep 2009.
Mining Retail Transaction Data for Targeting Customers with Headroom - A Case Study. MVS Shashanka, M Giering.
Artificial Intelligence Applications and Innovations, Greece, April 2009.
Probabilistic Factorization of Non-Negative Data with Entropic Co-occurrence Constraints. P Smaragdis, MVS Shashanka, B Raj, GJ Mysore.
Intl Conf on Independent Component Analysis, Brazil, March 2009.
Sparse and Shift-Invariant Feature Extraction from Non-Negative Data. P Smaragdis, B Raj, MVS Shashanka.
IEEE Intl Conf on Acoustics, Speech and Signal Processing, Las Vegas, Nevada, Apr 2008.
Sparse Overcomplete Latent Variable Decomposition of Counts Data. MVS Shashanka, B Raj, P Smaragdis.
Neural Information Processing Systems Conference (NIPS), Vancouver, Canada, Dec 2007.
Privacy-Preserving Musical Database Matching. MVS Shashanka, P Smaragdis.
IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, New Paltz, New York, Oct 2007.
Supervised and Semi-Supervised Separation of Sounds from Single-Channel Mixtures. P Smaragdis, B Raj, MVS Shashanka.
Intl Conf on Independent Component Analysis, London, UK, Sep 2007.
Sparse Overcomplete Decomposition
for Single Channel Speaker Separation. MVS Shashanka, B Raj, P Smaragdis. IEEE Intl Conf on Acoustics, Speech and Signal
Processing, Honolulu, Hawaii, April 2007.
A Framework for Secure Speech
Recognition. P Smaragdis, MVS Shashanka. IEEE Intl Conf on Acoustics,
Speech and Signal Processing, Honolulu, Hawaii, April
2007.
Bandwidth Expansion with a Polya
Urn Model. B Raj,
R Singh, MVS Shashanka, P Smaragdis. IEEE Intl
Conf on Acoustics, Speech and Signal Processing, Honolulu,
Hawaii, April 2007.
Separating a Foreground Singer from
Background Music. B Raj, P Smaragdis, MVS Shashanka, R Singh. Intl
Symposium on Frontiers of Research on Speech and Music (FRSM),
Mysore, India, Jan 2007.
A Probabilistic Latent Variable
Model for Acoustic Modeling. P Smaragdis, B Raj, MVS Shashanka. Workshop on Advances in Models for Acoustic
Processing, NIPS 2006.
Secure Sound Classification: Gaussian
Mixture Models.*
MVS Shashanka, P Smaragdis. IEEE Intl. Conf. on
Acoustics, Speech and Signal Processing, Toulouse, France, May
2006. *Finalist in the Student Paper
Contest.
Latent Dirichlet Decomposition for Single Channel Speaker Separation. B Raj, MVS Shashanka, P Smaragdis.
IEEE Intl. Conf. on Acoustics, Speech and Signal Processing, Toulouse, France, May
2006.
Optimal Multi-Channel Data Allocation
with Flat Broadcast per Channel.* AA Bertossi, MC Pinotti, S Ramaprasad,
R Rizzi, MVS Shashanka.
Intl. Parallel and Distributed Processing Symposium, Santa Fe, USA, Apr
2004. *Authors listed in alphabetical order.
A Characterisation of Optimal Channel Assignments for Wireless
Networks Modelled as Cellular and Square Grids. MVS Shashanka, A Pati, AM Shende.
Intl. Parallel and Distributed Processing Symposium, Nice, France,
Apr 2003.
Channel Assignment for Wireless
Networks Modelled as d-Dimensional Square Grids. A Dubhashi, MVS Shashanka, A Pati, S Ramaprasad, AM Shende.
Intl. Workshop on Distributed Computing, Kolkata, India, Dec
2002.
Posters / Abstracts
Mining Retail Data for Targeting Customers with Headroom. Madhu Shashanka, Michael Giering.
NYAS Annual Machine Learning Symposium, New York, NY, Oct 2008.
Separating and Understanding a Talker from a Mixture in Reverberant Spaces. BG Shinn-Cunningham, S Bressler, MVS Shashanka.
152nd meeting of the Acoustical Society of America (ASA), Honolulu, Hawaii, Dec 2006.
The Role of Fundamental Frequency in Segregating and Understanding a Talker Competing with Another Talker in a Reverberant Setting. MVS Shashanka, BG Shinn-Cunningham, S Nasser.
29th midwinter meeting of the Association for Research in Otolaryngology (ARO), Baltimore, Feb 2006.
Effects of Reverberant Energy on
Statistics of Speech. MVS Shashanka, BG Shinn-Cunningham, M Cooke.
Workshop on Speech Separation and Comprehension in Complex Acoustic Environments, Montreal, Canada, Nov 2004.
Probabilistic Latent Variable Model for Sparse Decompositions of Non-negative Data. MVS Shashanka, B Raj, P Smaragdis.
Unpublished Draft, March 2007.
Optimal Multi-Channel Data Allocation
with Flat Broadcast per Channel.* AA Bertossi, MC Pinotti, S Ramaprasad,
R Rizzi, MVS Shashanka.
Technical Report DIT-02-0079, University of Trento, 2003. *Authors listed alphabetically.
PATENTS
Method and System for Matching Audio Recordings. P Smaragdis and MVS Shashanka, USA Patent Application filed (27 Aug 2007).
System and Method for Recognizing Speech Securely. P Smaragdis and MVS Shashanka, USA Patent Application filed (16 Jan 2007).
Secure Classification of Data with Gaussian Distributions. MVS Shashanka and P Smaragdis, USA
Patent 7,526,084 (filed 02 Sep 2005, issued 28 Apr 2009).
TEACHING EXPERIENCE
Spring 2005
Boston University, Boston, MA
CN550 - Computational Models of Recognition, Memory and Attention
RESEARCH EXPERIENCE
Mar 2010 - Present
United Technologies Research Center, East Hartford, CT
Staff Scientist. Decision Support and Machine Intelligence.
Oct 2007 - Feb 2010
Mars Incorporated, Mount Olive, NJ
Research Scientist. I was part of a multi-disciplinary central business research group and internal think-tank that provided applied long-term research for Mars, Inc. My focus was machine learning, data mining and applied math modeling.
Summer 2006 - August 2007
Mitsubishi Electric Research Labs, Cambridge, MA
Research Intern. Worked on probabilistic models and algorithms for separating sources from single-channel audio recordings.
2004 - August 2007
Boston University, Boston, MA
Graduate Research Assistant in the Auditory Neuroscience Lab. Designed and conducted experiments to understand human auditory processing in natural reverberant environments, worked on psychoacoustics and developed computational/mathematical models of auditory perception.
Summer 2005
Mitsubishi Electric Research Labs, Cambridge, MA
Research Intern. Developed algorithms for secure classification between two parties such that private data is not exchanged; worked on sound separation algorithms.
Spring 2003
University of Trento, Trento, Italy
Research Fellow in the Department of Informatics and Telecommunication. Conducted research in the area of network algorithms. Developed efficient algorithms for optimally allocating data over multiple channels for broadcast such that average client waiting time is minimized.
Summer 2002
Indian Institute of Science, Bangalore, India
Summer Research Fellow in the Department of Electrical Engineering. Worked on applying learning automata to the problem of channel assignment in wireless networks.
Summer 2001
Bhabha Atomic Research Centre, Mumbai, India
Summer Intern in the Ultrasonics Instrumentation Section, Electronics Division. Developed a software module using Visual Basic for better visualization of scan data for an ultrasonic imaging system.
ENTREPRENEURIAL EXPERIENCE
Fall 2005
Biomimetic Systems, Boston, MA and Boston University, Boston, MA
Completed technology and venture assessments for the Boston University startup Biomimetic Systems.
TALKS / PRESENTATIONS
Probabilistic Latent Variable Models. United Technologies Research Center, Hartford, CT, 04 Nov 2009.
Simplex Decompositions - Techniques for Nonnegative Data Analysis and Extensions. United Technologies Research Center, Hartford, CT, 03 Sep 2009.
Simplex Decompositions. Icosystem Corporation, Boston, MA, 14 Aug 2009.
Latent Variable Framework for Modeling and Separating Single-Channel Acoustic Sources. Cognitive and Neural Systems Dissertation Defense, Boston University, Boston, MA, 17 Aug 2007.
Latent Variable Decomposition - Models and Applications (with Dr. Bhiksha Raj). MIT Brains and Machines Seminar, Cambridge, MA, 16 May 2007.
Probabilistic Latent Semantic Analysis. Catalyst Group - Mars Inc., Brussels, Belgium, 02 May 2007.
Probabilistic Models for Single-Channel Audio Processing. East Bay Institute of Research and Education, Martinez, CA, 26 Apr 2007.
Probabilistic Models for Single-Channel Audio Processing. Boston University Hearing Research Center Seminar, Boston, MA, 19 Jan 2007.
Probabilistic Models for Acoustic Processing. Mitsubishi Electric Research Laboratories,
Cambridge, MA, 21 Nov 2006.
SCHOLARSHIPS AND AWARDS
Finalist, Best Student Paper Award, Intl. Conf. on Acoustics, Speech and Signal Processing (ICASSP) 2006
Invited Outstanding Young Scientist, Workshop on Speech Separation and Comprehension in Complex Acoustic Environments (2004)
Arts and Science Dean's Fellowship, Boston University (2003 - 2004)
Research Fellowship Awardee (top 3% in India), JN Center for Advanced Scientific Research and Govt. of India (2002)
Merit Scholarship, Birla Institute of Technology and Science - awarded to top 10 students across majors in order of merit (1999 - 2002)
National Talent Search Scholarship, Govt. of India (1997 - 2003)
National top 1% among 17846 candidates, National Standard Exam in
Physics, India (1999)
Ranked 7th in Pre University Examinations (out of 319,028 students), State of Karnataka,
India (1999)
Ranked 1 in the State of Karnataka, National Talent Search Exam,
India (1997)
Ranked 16 in Secondary School Examinations (out of 471,307 students), State of Karnataka,
India (1997)
Ranked 6 in Regional Maths Olympiad, State of Karnataka, India (1997)
Ranked 13 in Regional Maths Olympiad, State of Karnataka, India (1996)
Ranked 12 in Inter-State Talent Tests in Maths, India (1996)
PROFESSIONAL ACTIVITIES & MEMBERSHIPS
Member, Institute of Electrical and Electronics Engineers (IEEE)
Scientific Committee Member, Workshop on Statistical and Perceptual Audition (SAPA) 2008
Technical Program Committee Member, Intl. Conf. on Multimedia and Expo (ICME) Multimedia Content Analysis Track 2007
Invited Participant, Workshop on New Ideas in Hearing, Paris, 2006
Invited Reviewer for IEEE Transactions on Signal Processing; IEEE Transactions on Audio, Speech and Language Processing; Neural Information Processing Systems Conference (NIPS); International Conference on Multimedia and Expo (ICME); Workshop on Statistical and Perceptual Audition
GRADUATE COURSES
Hearing
Neural and Computational Models of Speech and Hearing
Neural Coding and Perception of Sound (Harvard/MIT Speech and Hearing Program)
Psychoacoustics
Cognitive and Neural Systems
Computational Neuroscience (Introductory and Advanced)
