Reinforcement Learning with the Use of Costly Features.
Robby GoetschalckxScott SannerKurt DriessensPublished in: ECAI (2008)
Keyphrases
- reinforcement learning
- feature vectors
- function approximation
- markov decision processes
- classification accuracy
- low level
- data sets
- error prone
- structural information
- supervised learning
- image features
- prior knowledge
- object recognition
- learning algorithm
- machine learning
- learning process
- feature set
- false positives
- image sequences
- real time
- specific features