Deep Q-learning using redundant outputs in visual doom.
Hyun-Soo ParkKyung-Joong KimPublished in: CIG (2016)
Keyphrases
- reinforcement learning
- visual information
- cooperative
- state space
- function approximation
- visual features
- data sets
- visual perception
- learning rate
- dynamic programming
- low level
- high level
- computer vision
- learning algorithm
- deep learning
- real time
- temporal difference learning
- visual analysis
- visual appearance
- visual search
- image classification