Fig. 8
The normalized loss comparison for each CIMT approach in the (a) Essays and (b) Kaggle Dataset. Generally, the proposed FBCE-based approaches demonstrate faster convergence particularly in balanced Kaggle dataset.
Fig. 8 | Scientific Reports
The normalized loss comparison for each CIMT approach in the (a) Essays and (b) Kaggle Dataset. Generally, the proposed FBCE-based approaches demonstrate faster convergence particularly in balanced Kaggle dataset.
