Basics of few-shot learning with optimization-based meta-learning
Many machine learning models (particularly deep neural nets) require extensive training data. The idea of few-shot learning is to find ways to build models t...
Posts by year
2021
Understanding binary classifier model structure based on Shapley feature interaction patterns
I have written a full version of this article, available on bioRxiv. A copy of which is presented below.
Expectation-maximization in general and for Gaussian mixtures
In statistical inference, we want to find what is the best model parameters given the observed data. In the frequentist view, this is about maximizing the li...
Vaccine efficacies (with COVID-19 examples) - Bayesian posterior VE/CI calculations
It was interesting to see that Pfizer/BioNTech used a Bayesian analysis of the primary endpoint for its COVID-19 vaccine. There are a few posts on this alrea...
Gibbs sampling of multivariate probability distributions
This is a continuation of a previous article I have written on Bayesian inference using Markov chain Monte Carlo (MCMC). Here we will extend to multivariate ...
Bayesian inference using Markov Chain Monte Carlo with Python (from scratch and with PyMC3)
A guide to Bayesian inference using Markov Chain Monte Carlo (Metropolis-Hastings algorithm) with python examples, and exploration of different data size/par...
Monte Carlo integration in Python over univariate and multivariate functions
Monte Carlo integration is a basic Monte Carlo method for numerically estimating the integration of a function \(f(x)\). We will discuss here the theory alon...
2020
Getting started with Gaussian process regression modeling
Gaussian processing (GP) is quite a useful technique that enables a non-parametric Bayesian approach to modeling. It has wide applicability in areas such as ...