Dinoj Surendran

This page is a placeholder while I move my pages from my old University of Chicago site to Microsoft Research.

I am a Zimbrindian at Microsoft Research's Next Media Group working on data curation for the WorldWide Telescope.

I got my doctorate in computer science in December 2007 from the University of Chicago, where I worked in the intersection of computational linguistics and applied machine learning - quantifying the importance of phonological contrasts and investigating voice quality as a source of new acoustic information for recognizing tones in Mandarin Chinese. My advisor was Gina-Anne Levow, and the other members of my committee were Partha Niyogi and Chilin Shih. Here's my Ph.D. thesis (and data), which considers the following questions:

• If you are tone-deaf, how much information would you lose in trying to understand Mandarin? (Answer: about the same as if all vowels sounded alike.)
• Is voice quality helpful with tone recognition? (Not all measures are; the energy in various frequency bands does help with the recognition of neutral and low tone.)
• Are strongly articulated syllables easier to recognize? (In news broadcast speech, yes, but not by much.)

While in graduate school, I was part of what was fondly described as a "guerrilla astronomy visualization group" called COSMUS with some astronomy/outreach types at the Kavli Institute for Cosmological Physics, the University of Chicago Astronomy Department, and the Adler Planetarium & Astronomy Museum . We created stereo 3d interactive animations of the large scale structure of the universe, cosmic ray showers, and other astronomy datasets. These have since been used as visuals at a couple of museums, a NASA hi-def documentary, several classrooms, and Astronomy and Science magazines (see image to right).

My other hobby in grad school was using astronomy visualization packages to visualize the results of machine learning algorithms.

Wondering what Zimbrindian means? Ethnically I'm Indian, culturally I'm British, and I grew up in Zimbabwe.

I enjoy table tennis, reading Bujold, cycling very very slowly on flat ground, and pointing out the cloud that accompanies every silver lining.

World Time Zones : Seattle Area Weather

Publications

• Can Voice Quality help Mandarin Tone Recognition? by Dinoj Surendran and Gina-Anne Levow, Proceedings of ICASSP 2008.
  
  
• Dialog Act Tagging with Support Vector Machines and Hidden Markov Models, Dinoj Surendran and Gina-Anne Levow, Proceedings of Interspeech 2006.
  
  
• Interactive Glyph-Supplemented Scatterplots for Visualizing Algorithm Results. Dinoj Surendran. (Interactive Demo.) Companion Volume to the Proceedings of HLT-NAACL 2006. [Reviewer Comments] [Project Page]
  
  
• "Quantifying the functional load of phonemic oppositions, distinctive features, and suprasegmentals", Dinoj Surendran and Partha Niyogi, book chapter in Nedergaard Thomsen, Ole (ed. fc.) Current trends in the theory of linguistic change. In commemoration of Eugenio Coseriu (1921-2002). Amsterdam & Philadelphia: Benjamins.
  
  
• Tone Recognition in Mandarin using Focus. Dinoj Surendran, Gina-Anne Levow, Yi Xu (Phonetics Department, UCL). Proceedings of the 9th European Conference of Speech Communication and Technology (Interspeech/ICSLP 2005)
  
  
• Visualizing high dimensional datasets with Partiview. Dinoj Surendran and Stuart Levy (NCSA, UIUC). Poster Proceedings of the Tenth IEEE Conference on Information Visualization, Austin, TX, 2004.
  
  
• The functional load of tone in Mandarin is as high as that of vowels. Dinoj Surendran and Gina-Anne Levow. Proceedings of Speech Prosody 2004, Nara, Japan, pp. 99-102.

Non-First-Author Publications

• "Title Classification of Statement and Question Intonations in Mandarin", Fang Liu, Dinoj Surendran, and Yi Xu. Proceedings of Speech Prosody 2006, Dresden, Germany.
  
  
• Estimation of Stellar Atmospheric Parameters from SDSS ugriz Photometry. Y. Lee, T. C. Beers, C. Bailer-Jones, H. J. Newberg, M. Subbarao, D. Surendran. Proceedings of the 207th Meeting of the Astrononomical Association of America, January 2006.
  
  
• Articulatory complexity, ambient frequency and functional load as predictors of consonant development in children.. Stephanie Stokes (Dept of Psychology, University of Reading) and Dinoj Surendran. Journal of Speech and Hearing Research 48(3): 577-91, June 2005.

Technical Reports

All are from the University of Chicago Computer Science Department unless otherwise stated.

• Local Rhyme-based Acoustic Features for Mandarin Tone Recognition. Dinoj Surendran and Gina-Anne Levow (TR-2006-05)

• Additional Cues for Mandarin Tone Recognition, Dinoj Surendran and Gina-Anne Levow (TR-2006-04)

• Measuring the usefulness (functional load) of phonological contrasts. (PDF version). Dinoj Surendran and Partha Niyogi (TR-2003-12)

Non-research publications are listed on the Writing page.

Presentations

Here are presentations I've been involved in. I was the one doing the talking, unless otherwise mentioned. I need to add the WWT-related ones.

• May 2005: "Automating Visualization", at the workshop "Visualization of Astrophysical Data: Bringing Together Science, Art, & Education", Kavli Institute for Cosmological Physics, Chicago, IL.

• January 2005: "3d Visualizations of Current Astrophysical Research for Public Explorations and for the Classroom". Randy Landsberg, Mark SubbaRao and Dinoj Surendran. (Presented by Randy and Mark.) Presented at the Gadgets and Gizmos workshop at the 205th Meeting of the American Astronomical Society, San Diego, CA.

• December 2004: "Partiview: A Tool to Visually Analyze the Perfomance of Dimensionality Reduction and Clustering Algorithms". Dinoj Surendran and Stuart Levy (NCSA, UIUC). Demo at NIPS 2004.

• October 2004: "Visualizing High-dimensional datasets with Partiview". Dinoj Surendran and Stuart Levy. Joint presentation of an interactive poster at this year's Information Visualization conference in Austin, TX.

• October 2004: "Low-cost Stereo Visualization of Astronomical and Astrophysical Data". Randy Landsberg, Dinoj Surendran, Mark SubbaRao. (Presented by Randy and Dinoj.) Demo at the 4th Annual Great Lakes Planetarium Association Conference, Detroit, MI.

• September 2004: "Making Pretty Pictures for Machine Learning". Dinoj Surendran. Presentation at the Student Session at the Machine Learning Summer School, Berder Island, France.

• July 2004: "Visualizing the Sloan Digital Sky Survey in 3d Stereo". Mark SubbaRao and Dinoj Surendran. Joint presentation at the Physics Education and Outreach Workshop at the Aspen Center for Physics, Aspen, CO.

• March 2004: Cosmus: Virtual 3D Cosmology in Public Science Museums. "VR for Public Consumption" workshop at Virtual Reality 2004. Stephanie Andrews (presenter), Dinoj Surendran, Randall H. Landsberg, Eric Jojola, Leo Kadanoff, Ronen Mir, Joi Podgorny, Daniela Rosner, Mark SubbaRao.

• March 2004: "The functional load of tone in Mandarin is as high as that of vowels". Dinoj Surendran and Gina Levow. Poster at Speech Prosody 2004. Nara, Japan.

• November 2003. "The Quantification of Functional Load". Masters Thesis Presentation.

• August 2003: "Quantifying the Importance of Functional Load." Dinoj Surendran and Partha Niyogi. Paper presented at the 16-th Conference for Historical Linguistics, Copenhagen, Denmark.

• March 2002. "Language does not have to evolve optimally" Dinoj Surendran and Partha Niyogi. Fourth International Conference on Language Evolution, Harvard University, Cambridge MA.