In Reinforcement Learning, one way to solve finite MDPs is to use dynamic programming. Policy Evaluation (of the value functions) It refers to the iterative

Finite MDP is the formal problem definition that we try to solve in most of the reinforcement learning problem. Definition Finite MDP is a classical

Pre-requisite: some understanding of reinforcement learning. If not, you can start from Reinforcement Learning Primer Goal Let’s analyze this in the classic Multi-Armed Bandit problem using

Reinforcement learning is going to be “the next big thing” in machine learning after 2022, so let’s understand some basic on how it works. Agent:

The positional encoding of transformer was a detail added in Attention Is All You Need. When I first saw this, I thought “why is the

The KL (Kullback–Leibler) Divergence and JS (Jensen-Shanon) Divergence are ways to measure the distance (similarity) between two distributions P and Q. I will try to

Background In 2022, if you are not new to NLP (Natural Language Processing), you should have heard of BERT (Bidirectional Encoder Representations from Transforms). It’s

What are Chi-squared tests for? Compare an expected (hypothesized) categorical distribution vs an observed (sampled) categorical distribution Note that the distribution must be categorical (ie.

Goal Gradient descent seems to work fine for finding the local maxima and minima of a function, and Lagrange Multiplier helps to find the local

What are eigenvalues and eigenvectors? We have to answer both concept together since they are closely related? Given a square matrix A If we find