List of posts

• Random thoughts on ChatGPT
• Custom training loops with Pytorch
• Applications of autoencoders
• The joy of not google'ing: Short to long stick ratio in broken rods
• The expectation-maximization algorithm - Part 1
• Acquisition functions in Bayesian Optimization
• Bayesian optimization for hyperparameter tuning
• Longest substring with non-repeating characters
• Decision Trees: Gini index vs entropy
• An introduction to Gaussian Processes
• Alternating direction method of multipliers and Robust PCA
• Principal Component Analysis limitations and how to overcome them
• Why is normal distribution so ubiquitous?
• A list of machine-learning questions for interviews
• The encoder-decoder model as a dimensionality reduction technique
• Python decorators and the tf.function
• Probabilistic regression with Tensorflow
• How to implement a Naive Bayes classifier with Tensorflow
• Trainable probability distributions with Tensorflow
• Custom training loops and subclassing with Tensorflow
• Computing improper integrals with Cauchy's residue theorem
• A gentle introduction to kernel density estimation
• An almost one-liner to construct the Mandelbrot set with Mathematica
• Have you ever heard of anyone who became ill with Covid-19?
• Principal component analysis with Lagrange multiplier
• GitHub analytics with Mathematica
• Benford's law and COVID-19 conspiracy theories
• Bayesian connection to LASSO and ridge regression
• How to sum a divergent series with Pade approximation
• A simple example of perturbation theory
• How to convert DICOM transfer syntax
• Normality tests in statistics
• Common misconceptions about exponential growth
• Bayes theorem and likelihood ratios for diagnostic tests
• Dual spaces, dual vectors and dual basis
• The birthday paradox, factorial approximation and Laplace's method
• How to derive the Riemann curvature tensor
• Norms and machine learning
• What does it really mean for a machine to learn mathematics?
• Arcsin transformation gone wrong
• The meaning of curl operator
• Energy considerations for training deep neural networks
• Gradient descent
• Adversarial attacks on neural networks