-
Reinforcement Learning — From Reward-Penalty Rules to Q-Learning
A structured walkthrough of core RL concepts, from the reward-penalty weight update rule to Temporal Difference learning, SARSA, and Q-learning.
-
Unsupervised Learning — Neural Networks vs. Classical Machine Learning
A comparative review of unsupervised learning techniques across neural network and classical ML perspectives, from PCA and SOM to Autoencoders and Diffusion Models.
-
Supervised Recurrent Networks and the GRU to Vanishing Gradients
A deep dive into how Recurrent Neural Networks handle sequential data
-
EM Algorithm for Gaussian Mixture Models
Implementing GMM from scratch reminded me of the mathematical elegance behind classical statistical models — and why EM algorithm is an optimization masterpiece.