Colour versus Shape Goal Misgeneralization in Reinforcement Learning: A Case Study.
Karolis RamanauskasÖzgür SimsekPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- function approximation
- shape model
- multi agent
- state space
- shape descriptors
- case study
- temporal difference
- model free
- machine learning
- shape change
- shape recognition
- shape representation
- active contours
- neural network
- shape analysis
- vector field
- shape features
- dynamical systems
- d mesh
- transfer learning
- reinforcement learning algorithms
- shape retrieval
- learning algorithm