DeepLine: AutoML Tool for Pipelines Generation using Deep Reinforcement Learning and Hierarchical Actions Filtering.
Yuval HeffetzRoman VainshteinGilad KatzLior RokachPublished in: KDD (2020)
Keyphrases
- reinforcement learning
- action selection
- perceptual aliasing
- partially observable domains
- function approximation
- state space
- hierarchical structure
- machine learning
- decision theoretic
- state and action spaces
- hierarchical reinforcement learning
- reasoning about actions
- partially observable
- reward function
- model free
- plan recognition
- unsupervised learning
- information filtering
- temporal difference
- software tools
- learning capabilities
- generation process
- hierarchical clustering
- learning tasks
- markov decision processes
- learning process
- learning algorithm