A Robust Quantile Huber Loss with Interpretable Parameter Adjustment in Distributional Reinforcement Learning.
Parvin MalekzadehKonstantinos N. PlataniotisZissis PoulosZeyu WangPublished in: ICASSP (2024)
Keyphrases
- parameter adjustment
- reinforcement learning
- robust statistics
- robust estimation
- neural network
- function approximation
- information systems
- learning environment
- least squares
- computationally efficient
- temporal difference
- data sets
- bayesian networks
- multi agent systems
- dynamic programming
- optimal policy
- parameter tuning
- learning algorithm
- machine learning