CabSense

Fri, 9 Apr 2010 10:39:45 -0400

Over at Sense Networks, we just launched a new app for iPhone/Android to help New Yorkers find the best street corners to hail a cab. It uses using machine learning algorithms trained on over 90 million taxi trips, and has been featured in the NYTimes, The Wall Street Journal, Lifehacker, Brooklyn Ink (w/ video), and AM New York among others. For more information go to cabsense.com our check us out on twitter @cabsense.

Buddhabrot

Tue, 30 Dec 2008 14:12:44 -0500

Here is a large 50 megapixel rendering of the famous buddhabrot fractal. The code for the buddhabrot is adapted from Paul Bourke.

Underwater Bubble Ring

Wed, 15 Oct 2008 14:31:42 -0400

I used an underwater case for my canon G9 to capture this image of an underwater bubble ring that I blew in my pool this summer.

Sense Networks in NY Times

Thu, 26 Jun 2008 16:21:01 -0400

The startup I work for was just written up in The New York Times business section. It’s a great article talking about our Macrosense and Citysense products. Check out our website for more information. I made many of the visualizations on the website in processing and matlab.

Speech-Enabled Avatar Demo

Tue, 6 May 2008 17:30:53 -0400

The Columbia Computer Vision Lab is working on being able to create an avatar of someone from a single image that can speak in a somewhat realistic manner. This one happens to bear my likeness. I think it is awesome... and a little creepy.

Tze Go Ban Dance 360°

Mon, 28 Apr 2008 20:49:53 -0400

I finally figured out how to properly capture video with a one shot 360° camera, and the result is pretty interesting. This is a joint project with the Columbia Computer Vision Laboratory and the Tze Chun Dance Company. Check it out...

Conway’s Game of Life

Mon, 24 Mar 2008 14:31:06 -0400

Conway’s Game of Life is amazing. From incredibly simple rules and initial conditions, so many complex and rich patterns emerge. I decided to write a quick version in processing: Game of Life.

Also, I found this amazing implementation of a Turing Machine in the Game of Life. I recommend downloading Golly and checking it out. I knew that on a computational level Life is equivalent to a universal turing machine, but there is something about actually seeing it in action.

Huge Sand Spiral

Tue, 25 Sep 2007 01:25:45 -0400

What can I say... I like sand castles. This might be one of my finest...

More Flowers Pictures

Tue, 13 Mar 2007 21:54:35 -0400

Here are some great pictures of flowers I took while I was down in DC for the weekend... Flowers 2.

MVE

Thu, 21 Sep 2006 02:55:27 -0400

Minimum Volume Embedding (MVE) is an algorithm for non-linear dimensionality reduction that uses semidefinite programming (SDP) and matrix factorization to find a low-dimensional embedding that preserves local distances between points while representing the dataset in many fewer dimensions. MVE follows an approach similar to algorithms such as Semidefinite Embedding (SDE), in that it learns a kernel matrix using an SDP before applying Kernel Principal Component Analysis. However, the objective function for MVE directly optimizes the eigenspectrum of the data to preserve as much of its energy as possible within the few dimensions available to the embedding. Simultaneously, remaining eigenspectrum energy is minimized in directions orthogonal to the embedding thereby keeping data in a so-called minimum volume manifold. We show how MVE improves upon SDE in terms of the volume of the preserved embedding and the resulting eigenspectrum, producing better visualizations for a variety of synthetic and real-world datasets, including simple toy examples, face images, handwritten digits, phylogenetic trees, and social networks.

Learning from a Visual Folksonomy: Automatically Annotating Images from Flickr

Wed, 3 May 2006 02:13:14 -0400

Recently, a large visual dataset has emerged from a web-based photo service called Flickr which utilizes the organizational power of folksonomy to label a tremendous amount of visual data. Flickr users upload snapshots from their digital cameras to the web, and if marked as public, the community annotates these images with descriptive tags. Can this large collective labeling effort be used to train a computer to annotate images? What concepts are we able to train a computer to visually identify?

This project uses a simple crawler to download photos from Flickr labeled with a certain tag, and then extracts color and texture features from these images so that they can be used to train a classifier, such as a Support Vector Machine (SVM). By automating this process of downloading images, extracting features, training, and testing, we are able to apply our system to many different tags and see which tags correspond to identifiable visual features. We have found that the system performs relatively well annotating images with one label, selected from a small vocabulary, for images belonging to concepts with distinct color and texture features. (Full paper found here)

Visualizing Folksonomies using Machine Learning Algorithms

Wed, 19 Oct 2005 02:57:12 -0400

This paper, written for my Adv. Machine Learning class, investigates using Semidefinite Embedding (SDE) to visualize data collected from a folksonomy. The del.icio.us social bookmarking service is a perfect example of a folksonomy; a community of users label websites with descriptive tags. Each tag exists in a high-dimensional space corresponding to the frequency of use of that tag among all the users of the system. We are motivated by the following question: can we find a simple low-dimensional structure for these tags that captures the significant relationships inherent in the data? In this paper we explore Semidefinite Embedding, an algorithm for non-linear dimensionality reduction, and its application to visualizing folksonomic systems, focusing on the effects of specifying different levels of connectivity for the data and the heuristics which can be used to find the best parameters for the algorithm. (Full paper found here)

Utilizing Folksonomy: Similarity Metadata from the Del.icio.us System

Wed, 12 Oct 2005 03:01:22 -0400

Traditionally, metadata is thought of simply as keywords that describe some content, and while the primary aim of folksonomic systems like the Del.icio.us bookmarking tool is to produce these keywords, a richer set of metadata is also produced. Because these keywords are now contributed from many different individuals and aggregated, useful information comes not only from the keyword itself but also from the information about who contributed to labeling the content with that keyword. This idea can be broadened to a general framework for producing a new layer of metadata: similarity between concepts. By analyzing the distributions of how users apply tags, how tags are applied to links, and how users pick content, we should be able to calculate the "distance" between tags, users, and content. This "distance" metric could then be used to construct a more powerful tool for browsing content, allowing the user to specify a query made up of keywords, content, or even other users. Furthermore, this metadata can be condensed into a lower dimensional space and visualized in order to gain better insight into the relationships between the concepts themselves. (Full paper found here)

CUtunes Update

Wed, 21 Sep 2005 02:59:13 -0400

CUtunes is looking better after another semester of work. Here is the updated documentation, and some screenshots. Notable new features are user profile pages, flash-based visualizations of your musical neighborhood, and inteligent playlist creation in itunes, allowing the user to say make me a playlist that is like a specified list of musical artists and CUtunes users.

Estimating PI

Wed, 17 Mar 2004 01:34:05 -0500

Here is a flash visualization of Buffon’s Needle, a method for calculating PI based on the statistics of repeatedly dropping matchsticks on the ground.