Wave Function Collapse is a procedural generation algorithm which produces images by arranging a collection of tiles according to rules about which tiles may be adjacent to each other tile, and relatively how frequently each tile should appear. The algorithm maintains, for each pixel of the output image, a probability distribution of the tiles which may be placed there. It repeatedly chooses a pixel to “collapse” - choosing a tile to use for that pixel based on its distribution. WFC gets its name from quantum physics. The goal of this post is to build an intuition for how and why the WFC algorithm works.
While implementing a quick toy example of Crane and Sawhney's really great Monte Carlo Geometry Processing paper, the question arose about whether a quick function I grabbed from The Internet to equally distribute points on a sphere was correct or not. Since it's absolutely the crux of the method, this is an important question! This notebook performs a rather unscientific check for equal distribution of points on the surface of a sphere. It uses the first algorithm from MathWorld: Sphere Point Picking. Foll
GPUs are designed to do many things well, but drawing transparent 3D objects is not one of them. Opacity doesn't commute so that the order in which you draw surfaces makes a big difference. Of course simple additive blending does commute, but it's not really what we think of as "transparent objects". The simplest way to draw transparent objects is from back to front via the painter's algorithm. In this approach we sort geometry and draw only from back to front. This requires sorting triangles, which, in add
Hi Geeks, welcome to Part-3 of our Reinforcement Learning Series. In the last two blogs, we covered some basic concepts in RL and also studied the multi-armed bandit problem and its solution methods…
When the agent interacts with the environment, the sequence of experienced tuples can be highly correlated. The naive Q-Learning algorithm that learns from each of these experience tuples in…
In Q-Learning, we represent the Q-value as a table. However, in many real-world problems, there are enormous state and/or action spaces and tabular representation is insufficient. For instance…
This is a PyTorch implementation/tutorial of Deep Q Networks (DQN) from paper Playing Atari with Deep Reinforcement Learning. This includes dueling network architecture, a prioritized replay buffer and double-Q-network training.
In this article, we will try to understand where On-Policy learning, Off-policy learning and offline learning algorithms fundamentally differ. Though there is a fair amount of intimidating jargon in…