You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm new to the field of deep learning. If I want to emphasize some specific characteristics of the drug molecule in the process of training a model, such as bond length and bond angle. How do I change the code?
The text was updated successfully, but these errors were encountered:
Generally speaking, a model is trained to use all features, determining useful features on its own. Can you be more specific about emphasizing characteristics?
I mean if I have already realized in advance that certain features may play a crucial role in the training of the model. Is there any way to make my model also recognize that these features are important?
Best Regards
It depends on how do you define "importance". If the target is to make certain features easier to learn or to be separated, following the philosophy of PCA can be a good way--just scale up the variances of certain features.
I'm new to the field of deep learning. If I want to emphasize some specific characteristics of the drug molecule in the process of training a model, such as bond length and bond angle. How do I change the code?
The text was updated successfully, but these errors were encountered: