PROJECTS

	Can One Hear the Shape of a Trend? (aka Music of the (Twitter)Spheres) I sonified the process of a topic spreading from person to person on Twitter. The resulting notes represent the natural frequencies that characterize the “shape” of the evolving network of users participating in a topic.
	Information Diffusion on Twitter I tracked and visualized how topics spread on Twitter implicitly from person to person by using the timing of the tweets and the follow graph.
	Control of Underactuated Vehicular Traffic When self-driving robotic cars begin to share the road with regular cars, could we control the robotic cars to smooth out traffic jams? I did a numerical and theoretical study of vehicle traffic dynamics and control policies that smooth traffic even when controlling only a fraction of the vehicles.
	Structure Formation in Microtubule Networks Large-scale simulation of axonal microtubule network structure formation. Got to play wiht Monte-Carlo methods, sampling algorithms, large-scale collision detection, and network visualization.
	Principal Component Landscapes Renderings of principal components handwritten ‘2’s as landscapes
	MASLAB (MIT Mobile Autonomous Vision-Based Robotics Competition)
	Cortical Learning Algorithms at Numenta I worked on Numenta’s Cortical Learning Algorithms. Prototyped and evaluated machine learning algorithms based on the structure and function of the neocortex, specifically online unsupevised visual feature learning, and wrote a system for interactive 3D visualization of experiment runs.
	A Nonparametric Method for Early Detection of Twitter Trends
	Semi-Supervised Shape Classification with Manifold Regularization Shape classification using manifold regularization.
	Probabilistic Model of Signal Onset Detection Statistical simulations of signal onset detection in the presence of noise. Modeled detection as a random walk with drift crossing a threshold. Model explained two paradoxes in the human time perception literature.
	Computer Graphics and Animation Engine 3D computer graphics and animation engine in Java from scratch. Line drawing, curve and mesh generation, lighting and shading, ray-tracing, a hierarchical scene-graph and motion description language, and tooling to create scenes and animations.
	Recommending Charitable Action In Response to News Recommendation system to match news stories with relevant opportunities for donation or volunteering. Text parsing, stemming, part-of-speech tagging. Cosine similarity. Bayesian inference to parse news article text from raw HTML. Prototype helped secure a round of angel funding.

Publications

• 2010

  • Probabilistic Model of Onset Detection Explains Paradoxes in Human Time Perception. With Hakwan Lau, Dobromir Rahnev. Frontiers in Psychology. 1:37 2010.

Conferences

• 2013

  • A Latent Source Model for Nonparametric Time Series Classification. With George H. Chen and Devavrat Shah. Neural Information Processing Systems (NIPS) 2013. (to appear)

• 2012

  • A Nonparametric Method for Early Detection of Trending Topics. With Devavrat Shah. Workshop on Information and Decision in Social Networks 2012.

• 2010

  • Integrative image analysis of Drosophila in-situ hybridization data (Poster). With Anna Ayuso, Christopher A. Bristow, Charlie Frogner, Manolis Kellis, Tom Morgan, Tomaso Poggio, Lorenzo Rosasco. 2010 CSHL Meetiing on Automated Imaging and High-Throughput Phenotyping.

• 2009

  • Signal Detection-Theoretic Model Explains Libet’s Awareness of Intention and Attentional Prior Entry (Poster). With Matthew Davidson, Tashina Graves, Hakwan Lau, Dobromir Rahnev. ASSC 13. 2009.

Talks

• 2013

  • Implicit Information Diffusion Networks on Twitter. INFORMS 2013. Session on Models of Information Diffusion in Social Networks. (Invited)
    

• 2012
  • A Nonparametric Method for Early Detection of Trending Topics. (Joint work with Devavrat Shah). Workshop on Information and Decision in Social Networks 2012. November 9, 2012 at MIT.
  • Information Diffusion on Twitter. Guest lecture for the course Analyzing Big Data with Twitter at the UC Berkeley School of Information. October 25, 2012 at UC Berkeley.

Thesis

• Trend or No Trend: A Novel Nonparametric Method for Classifying Timeseries. MEng Thesis, Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology. 2012.

Advertisements