Struct CoreSetSelection 

pub struct CoreSetSelection { /* private fields */ }

Core-Set selection for diversity-based active learning.

Selects samples that maximize coverage of the feature space, ensuring diverse representation in the labeled set.

Algorithm

Uses greedy furthest-first traversal to select points that are maximally distant from already selected points.

Reference

• Sener, O., & Savarese, S. (2018). Active Learning for Convolutional Neural Networks: A Core-Set Approach. ICLR.

Implementations

impl CoreSetSelection

pub fn new() -> Self

Create a new core-set selector.

pub fn with_labeled(labeled_indices: Vec<usize>) -> Self

Create with initial labeled indices.

pub fn select(&self, embeddings: &[Vec<f32>], k: usize) -> Vec<usize>

Select k samples using furthest-first traversal.

Arguments

• embeddings - Feature embeddings for all samples
• k - Number of samples to select

Returns

Indices of selected samples.

pub fn diversity_score(&self, embeddings: &[Vec<f32>], indices: &[usize]) -> f32

Compute diversity score for a set of indices.

Higher score indicates more diverse selection.

Trait Implementations

impl Clone for CoreSetSelection

fn clone(&self) -> CoreSetSelection

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for CoreSetSelection

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for CoreSetSelection

fn default() -> Self

Returns the “default value” for a type.