daniel-at-world.blogspot.com
Daniel@World: Spectrogram Interest Points: Shazaam
http://daniel-at-world.blogspot.com/2014/11/spectrogram-interest-points-shazaam.html
Machine Learning, Data Science, Computer Vision. Tuesday, November 11, 2014. Spectrogram Interest Points: Shazaam. After some time, I decided to write another blog post. This time I want to talk about what one can do with local interest points in a spectrogram. An interest point in a spectrogram is a point of high magnitude in time and frequency. Normally we use points that are a local maximum in a small region. For indexing is extracting such interest points from the spectrogram and continues to. Buildi...
daniel-at-world.blogspot.com
Daniel@World: Thoughts on EM - Clustering: Hierarchical Agglomerative Clustering
http://daniel-at-world.blogspot.com/2013/01/thoughts-on-em-clustering-hierarchical.html
Machine Learning, Data Science, Computer Vision. Tuesday, January 22, 2013. Thoughts on EM - Clustering: Hierarchical Agglomerative Clustering. A common clustering algorithm for continues data is using Expectation-Maximisation. Of Gaussian Mixture Models. In short a Gaussian Mixture Model is a joint distribution of k- Gaussians. Each of them has a prior probability of picking that gaussian. The joint probability of that model is given as:. Location: Atlanta, GA, USA. Subscribe to: Post Comments (Atom).
daniel-at-world.blogspot.com
Daniel@World: A night in stockholm airport
http://daniel-at-world.blogspot.com/2014/08/a-night-in-stockholm.html
Machine Learning, Data Science, Computer Vision. Thursday, August 21, 2014. A night in stockholm airport. Traveling back to the US for the fall semester at tech, I have a stop for a night in Stockholm / Sweden. Since I am bored I will blog about it :D. I arrive in Sweden at 12AM August 21st and leave the following day at 10 PM:. I just bought a completely overpriced plate of swedish meet balls. It was a disappointing experience and I think the ones you get at Ikea are better. Furthermore,. Podcasts can l...
daniel-at-world.blogspot.com
Daniel@World: Integrating Weka's UI into your own code
http://daniel-at-world.blogspot.com/2015/05/integrating-wekas-ui-into-your-own-code.html
Machine Learning, Data Science, Computer Vision. Wednesday, May 13, 2015. Integrating Weka's UI into your own code. Is a machine learning toolkit written in Java. Despite it's capability to run as a stand alone. Choosing and configuring a Weka classifier, here a Support Vector Machine. After browsing a little through the Weka UI code I found two classes that can be used to. Open the editor to choose a classifier and to configure it. Once the window closes, the classifier. Null) { Classifier classifier = ...
daniel-at-world.blogspot.com
Daniel@World: ISWC@Zürich and other impressions: No math this time I promise :D
http://daniel-at-world.blogspot.com/2013/09/iswczurich-and-other-impressions-no.html
Machine Learning, Data Science, Computer Vision. Sunday, September 15, 2013. ISWC@Zürich and other impressions: No math this time I promise :D. So as I said we (a lot of tech folks and me) went to Zürich for a nameless workshop and the International Symposium on Wearable Computing. The first day off between the workshop and the conference we explored Zürich a "bit" by walking 15 km around Zürich and hiked a mountain nearby. If you are interested you can download. The highlight for a lot of Georgia Tech p...
daniel-at-world.blogspot.com
Daniel@World: K-Means works just fine
http://daniel-at-world.blogspot.com/2014/12/merry-christmas-from-self-taught.html
Machine Learning, Data Science, Computer Vision. Sunday, December 28, 2014. K-Means works just fine. It is interesting how powerful vector quantization can be. Since I like the quantization idea a lot, I think the following insight from a 2011 paper. Was notable: K-Means outperforms single layer neural nets for self taught learning. The task is to learn features of an image from small patches extracted in a sliding window. Features are extracted over these patches by:. In the hidden layer. Location: Oste...
daniel-at-world.blogspot.com
Daniel@World: Bayes Meal Planning and Reach for Friends
http://daniel-at-world.blogspot.com/2012/03/bayes-meal-planning-and-reach-for.html
Machine Learning, Data Science, Computer Vision. Tuesday, March 27, 2012. Bayes Meal Planning and Reach for Friends. It is a long time since my last post. I figured you might be wondering what I currently do and how it is studying at Georgia Tech. So I thought giving you a quick outlook on what is going on in one of my classes (yes in the U.S people have to take classes for their PhD). In my current graduate AI Class (Gatech CS 6601) we had to perform two mini research projects. See Figure 1 ). Since com...
daniel-at-world.blogspot.com
Daniel@World: ICASSP: Pattern Discovery in Dolphin Whistles
http://daniel-at-world.blogspot.com/2014/08/icassp-pattern-discovery-in-dolphin.html
Machine Learning, Data Science, Computer Vision. Friday, August 15, 2014. ICASSP: Pattern Discovery in Dolphin Whistles. Abstract of my ICASSP 2014. Paper on Dolphin Communication Mining:. Location: 27711 Osterholz-Scharmbeck, Germany. Subscribe to: Post Comments (Atom). I am a Data Scientist working at the social network Xing in Hamburg. My main interests are in Machine Learning and Data Analysis, Artificial Intelligence, Wearable Computing and Pattern Recognition. . View my complete profile.
daniel-at-world.blogspot.com
Daniel@World: Building a Gesture Recognizer - Part 1 Dynamic Time Warping
http://daniel-at-world.blogspot.com/2012/11/building-gesture-recognizer-part-1.html
Machine Learning, Data Science, Computer Vision. Thursday, November 1, 2012. Building a Gesture Recognizer - Part 1 Dynamic Time Warping. I recently noticed a lot of people want to do "something with gestures" but don't know how to start. So I thought writing a two part tutorial on how to recognize gestures. The first part will. Describe the problems of gesture recognition and will describe a very simple algorithm, while. And to build a simple, distance based gesture recognizer. Exactly look the same....
daniel-at-world.blogspot.com
Daniel@World: Baum Welch Initialisation: Flat Start vs Randomisation vs Viterbi Training
http://daniel-at-world.blogspot.com/2013/07/baum-welch-initialisation-flat-start-vs.html
Machine Learning, Data Science, Computer Vision. Sunday, July 14, 2013. Baum Welch Initialisation: Flat Start vs Randomisation vs Viterbi Training. As most of the readers will know Baum Welch is a parameter estimation Algorithm for Hidden Markov Models. Given a initial configuration of the HMM it will converge to a locally. Viterbi training is implemented in HTK3 in the tool HInit. First the data is uniformly segmented. Random Initialisation with Random Restarts. Location: Bremen, Germany. I am a Data Sc...
SOCIAL ENGAGEMENT