Minimizing the Negative Log-Likelihood, in English

 — 

Statistical underpinnings of the machine learning models we know and love. A walk through random variables, entropy, exponential family distributions, generalized linear models, maximum likelihood estimation, cross entropy, KL divergence, maximum a posteriori estimation and going "fully Bayesian."


Transfer Learning for Flight Delay Prediction via Variational Autoencoders

 — 

Autoencoding airports via variational autoencoders to improve flight delay prediction. Additionally, a principled look at variational inference itself and its connections to machine learning.


Deriving the Softmax from First Principles

 — 

Deriving the softmax from first conditional probabilistic principles, and how this framework extends naturally to define the softmax regression, conditional random fields, naive Bayes and hidden Markov models.


Approximating Implicit Matrix Factorization with Shallow Neural Networks

 — 

In this post, we look to beat the performance of Implicit Matrix Factorization on a recommendation task using 5 different neural network architectures.


Ordered Categorical GLMs for Product Feedback Scores

 — 

A follow-up to Erik Bernhardsson's post "More MCMC – Analyzing a small dataset with 1-5 ratings" using ordered categorical generalized linear models.


Intercausal Reasoning in Bayesian Networks

 — 

Simple intercausal reasoning on a 3-node Bayesian network.


Bayesian Inference via Simulated Annealing

 — 

A toy, hand-rolled Bayesian model, optimized via simulated annealing.


RescueTime Inference via the "Poor Man's Dirichlet"

 — 

Modeling a typical week of RescueTime data via an alternative take on the Dirichlet distribution.


Generating World Flags with Sparse Auto-Encoders

 — 

Hand-rolled sparse autoencoders to generate novel world flags.


Docker and Kaggle with Ernie and Bert

 — 

An introduction to what Docker is and why and how to use it for Kaggle.


© Will Wolf 2017

Powered by Pelican