Newbeeer / L_DMI

Code for NeurIPS 2019 Paper, "L_DMI: An Information-theoretic Noise-robust Loss Function"

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Confusion Matrix

Faldict opened this issue · comments

Hi,

Your code looks good! But why do you set confusion matrix rather simple? I don't think your noisy data is really noisy. Have you tried more sophisticated cases?

conf_matrix = torch.eye(10)

Hi,

Very sorry for the huge delay. I have add the class-independent noise-type to this repo and the original paper. Please refer to our camera-ready version https://newbeeer.github.io/newbeeer.github.io/assets/files/dmi.pdf, where we include both the class-dependent noise (in the Sec 5.2) and the class-independent noise (in Appendiex D).

If you are interested in any other specific cases, please let me know. I am glad to have a try.

Thanks!

Best,
Yilun