Mila

1. Uncetainty mapping in meta learning

2. Simple frame-work

3. Meta-learning

4. Heteroscadastic

5. Homoscadastic

6. Epistmic

7. Water-body mapping

8. Self training

9. Distillation

10. results

11. challenges

11.1. There is no extra information we are adding while predicting in a new unlabelled data, that may cause some amount of error, but can be taken as baseline for further study.