Non-Markovian Reward Modelling from Trajectory Labels via Interpretable Multiple Instance Learning.
Joseph EarlyTom BewleyChristine EversSarvapali D. RamchurnPublished in: CoRR (2022)
Keyphrases
- multiple instance learning
- multiple instance
- reinforcement learning
- image categorization
- class labels
- positive bags
- multi class
- supervised learning
- image annotation
- semi supervised
- semi supervised learning
- reward function
- object based image retrieval
- pairwise
- training data
- instance selection
- multi label
- image processing
- learning problems
- unlabeled data
- text mining
- training set
- image segmentation
- data sets