About I started writing this post because of the most common misconception that I have always found in Machine Learning, almost always causing some practical problems. But the truth is that I extended the post much more and ended up becoming also one about how… Read More »The most common mistake in Machine Learning: How to quantify uncertainty? Likelihood vs Confidence
Even the greatest can be wrong. In this project I tell part of my experience with an incredible Google paper, and how I improved it by correcting a simple fundamental mistake. This post shows how important it is to always go back to basics and… Read More »Lifetime Value Prediction: A Mode vs Mean problem
Machine learning is changing the world. One of the most impressive advances are generative models. Over time it will be increasingly difficult to differentiate a fake video, image or audio from a real one. Having fun applications, others very useful, and could even become dangerous… Read More »Conditional Generative Adversarial Neural Networks for Metal Artifact Reduction in Spine CT Images
At the beginning of the pandemic, while I was working at Las Condes Hospital, a project was developed together with Amazon Web Services Chile for the detection of COVID-19 pneumonia patterns in chest X-rays. Using weighted cost functions for unbalanced data, transfer learning and augmentation.… Read More »A Rapid COVID-19 Detection with AWS Chile