Reinforcement Learning (RL) is a framework to train an intelligent agent to take a sequence of actions in order to maximize a reward, given the feedback from a dynamic environment.

Diffusion is one of the most used models for image / video generation. This note covers some basics of diffusion models.

The previous post introduces the basics of shadow mapping. However, I didn’t mention one of the fundamental problems: aliasing. If we take a closer look at the shadow, we will find it jagged on the edge.

Shadow mapping is a technique to simulate shadows. The basic version could create hard shadows. The advanced versions could even create soft shadows, and look very real. This article walks through some basics, and the demo comes here.

Prefix Sum problem is to compute the sum of all the previous elements in an array. Specifically, exclusive prefix sum would compute all the strictly previous (self-exclusive) elements. For example,

• Able to
  • Answer sum query within interval.
  • Update within fixed-size array.
• Variant
  • Answer prefix query of min/max.
• Code

Recently I was following an online course on Convolutional Neural Networks (CNN) provided by Stanford. I find it a very nice hands-on material: slides and notes are easy to understand. Purely reading formulations can be confusing sometimes, but practicing experiments helps better understanding what the formulations and the symbols in them are expressing.