The Bitcoin Crash and How Nature Works

A few years after this post appeared, a paper appeared on the same topic, analyzing 4 different Bitcoin bubbles. On … More

SVM+ / LUPI: Learning Using Privileged Information

Recently Facebook hired Vapnik, the father of the Support Vector Machine (SVM) and the Vapnik-Chrevoniks Statistical Learning Theory. Lesser known, Vapnik … More

Machine Learning with Missing Labels Part 3: Experiments

In this series of posts we look at Transductive and SemiSupervised Learning–an old problem, a hard problem, and a fundamental problem Machine Learning. Unlike Deep … More

Machine Learning with Missing Labels Part 2: The UniverSVM

Ever wonder what Google DeepMind is up to? They just released a paper on Semi-Supervised learning with Deep Generative Models. What is Semi … More

Machine Learning with Missing Labels: Transductive SVMs

SVMs are great for building text classifiers–if you have a set of very high quality, labeled documents. Frequently, we just … More

Data Science Leadership

My talk from Berkeley is up on the new Calculation Consulting YouTube channel https://www.youtube.com/channel/UCaao8GHAVcRtSZdpObC4_Kg

Foundations: The Partition Function.

We are going to examine the Partition function that arises in Deep Learning methods like Restricted Boltzmann Machines. We take … More

Music Recommendations and the Logistic Metric Embedding

In this post, we are going to see how to build our own music recommender, using the Logistic Metric Embedding … More

A Ruby DSL Design Pattern for Distributed Computing

Frequently in my work in big data and machine learning, I need to run large calculations in parallel. There are … More

Causality, Correlation, and Brownian Motion

A recent question on Quora asked if machine learning could learn something from the Black Scholes model of Finance http://www.quora.com/Machine-Learning/Can-we-learn-any-lessons-from-the-Black-Scholes-solution-to-pricing-risk-to-machine-learning-algorithms-for-personalization-recommendation-algorithms … More