mixup: mixup Function

In makeyourownmaker/mixup: Create Convex Combinations of Pairs of Examples and their Labels for Data Augmentation

mixup Function

Description

This function enlarges training sets using linear interpolations of features and associated labels as described in https://arxiv.org/abs/1710.09412.

An R package inspired by mixup: Beyond Empirical Risk Minimization

Usage

mixup(x1, y1, alpha = 1, concat = FALSE, batch_size = NULL)

Arguments

x1	Original features
y1	Original labels
alpha	Hyperparameter specifying strength of interpolation
concat	Concatenate mixup data with original
batch_size	How many mixup values to produce

Details

The x1 and y1 parameters must be numeric and must have equal numbers of examples. Non-finite values are not permitted. Factors should be one-hot encoded.

For now, only binary classification is supported. Meaning y1 must contain only numeric 0 and 1 values.

Alpha values must be greater than or equal to zero. Alpha equal to zero specifies no interpolation.

The mixup function returns a two-element list containing interpolated x and y values. Optionally, the original values can be concatenated with the new values.

Value

A list containing interpolated x and y values and optionally the original values

Interpolation

This package enlarges training sets using linear interpolations of features and associated labels:

x' = λ * x_i + (1 - λ) * x_j, where x_i, x_j are raw input vectors

y' = λ * y_i + (1 - λ) * y_j, where y_i, y_j are one-hot label encodings

(x_i, y_i) and (x_j ,y_j) are two examples drawn at random from the training data, and λ \in [0, 1] with λ \sim Beta(α, α) for α \in (0, ∞). The mixup hyper-parameter α controls the strength of interpolation between feature-target pairs.

Examples

# Use builtin mtcars dataset with mtcars$am (automatic/manual) as binary target
data(mtcars)
mtcars.mix <- mixup(mtcars[, -9], mtcars$am)

makeyourownmaker/mixup documentation built on Sept. 14, 2022, 8:51 a.m.