Skip to content

Latest commit

 

History

History
20 lines (16 loc) · 513 Bytes

README.md

File metadata and controls

20 lines (16 loc) · 513 Bytes

ecg-loss

Keras-compatible Tensorflow implementation for ε-contaminated Gausssian distribution loss.

Usage

Can be used with Keras using the Tensorflow backend as follows:

from ecg_loss import get_ecg_loss_func

model.compile( 
    loss=get_ecg_loss_func(ecg_c=1.0, ecg_epsilon=0.1),
    optimizer="adam",
    metrics=['mse'])

References

J. Tukey, “A survey of sampling from contaminated distributions,”
Contributions to probability and statistics, vol. 2, pp. 448–485, 1960.