position_regularizer: Position-wise lasso penalty applied to convolutional filters...

Description Usage Arguments Value Author(s) See Also

Description

L1 regularization penalty to constrain individual filter weights with decaying strength nearing the center of tte filter.

Usage

1
position_regularizer(weight_matrix)

Arguments

weight_matrix

Filter weights.

Value

Lasso penalty.

Author(s)

Matthew Ploenzke, ploenzke@g.harvard.edu

See Also

filter_regularizer total_regularizer


mPloenzke/learnMotifs documentation built on May 27, 2019, 11:55 a.m.