Struct AdaptiveCurriculum 

pub struct AdaptiveCurriculum { /* private fields */ }

Adaptive curriculum that adjusts based on error class performance

Tracks accuracy per error class and increases difficulty for well-learned classes while maintaining focus on struggling classes.

Supports the CITL adaptive tier selection pattern.

Implementations

impl AdaptiveCurriculum

pub fn new() -> Self

Create new adaptive curriculum

pub fn tier_for_error(&self, error_code: &str, attempt: usize) -> usize

Get recommended tier for an error class

Based on the CITL select_tier() pattern

pub fn update_class(&mut self, error_code: &str, correct: bool)

Update accuracy for an error class

pub fn weight_for_class(&self, error_code: &str) -> f32

Get sample weight based on class rarity and accuracy

Long-tail (rare) errors get higher weights per Feldman (2020)

Trait Implementations

impl Clone for AdaptiveCurriculum

fn clone(&self) -> AdaptiveCurriculum

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl CurriculumScheduler for AdaptiveCurriculum

fn difficulty(&self) -> f32

Get the current difficulty level (0.0 = easiest, 1.0 = hardest)

fn tier(&self) -> usize

Get the current tier (for tiered training like CITL)

fn step(&mut self, _epoch: usize, accuracy: f32)

Advance the curriculum based on training progress

fn reset(&mut self)

Reset the curriculum to initial state

fn name(&self) -> &'static str

Name of the curriculum scheduler

fn sample_weight(&self, sample_difficulty: f32) -> f32

Get sample weight for a given difficulty score

fn include_sample(&self, sample_difficulty: f32) -> bool

Check if sample should be included at current difficulty

impl Debug for AdaptiveCurriculum

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

Formats the value using the given formatter.

impl Default for AdaptiveCurriculum

fn default() -> Self

Returns the “default value” for a type.