Multi-Modal Mutual Information (MuMMI) Training for Robust Self-Supervised Deep Reinforcement Learning.
Kaiqi ChenYong LeeHarold SohPublished in: CoRR (2021)
Keyphrases
- multi modal
- mutual information
- multi modality
- reinforcement learning
- information theoretic
- image registration
- learning algorithm
- humanoid robot
- audio visual
- similarity measure
- semantic concepts
- video search
- cross modal
- state space
- high dimensional
- multiple modalities
- fusing multiple
- image processing
- metadata
- computer vision