Deep Reinforcement Learning
Module 4

Visual Reinforcement Learning

Train agents directly from pixels using convolutional policies, frame stacking, and ViZDoom — from a stationary-enemy basic scenario to multi-target arena combat.

3 readings 1 lab
Readings
1
Visual RL: From Pixels to Policies
12 min
2
Convolutional Policies and Preprocessing
13 min
3
ViZDoom and 3D Visual RL
11 min
Labs
Visual RL with ViZDoom
75 min