Module ml 

Machine Learning Simulation Engine.

Provides deterministic, reproducible simulation of ML workflows using Popperian falsification methodology. Implements TPS principles:

• Jidoka: Stop-on-anomaly detection
• Heijunka: Load-balanced batch processing
• Kaizen: Continuous improvement via feedback

Example

use simular::domains::ml::{TrainingSimulation, TrainingConfig, AnomalyDetector};
use simular::engine::rng::SimRng;

let mut sim = TrainingSimulation::new(42);
let config = TrainingConfig::default();
// Training simulation would run here

Re-exports

pub use jidoka::*;

pub use multi_turn::*;

pub use prediction::*;

Modules

jidoka

multi_turn

prediction

Structs

AnomalyDetector

Anomaly detector for Jidoka-style training quality gates.

RollingStats

Rolling statistics for anomaly detection.

TrainingConfig

Training hyperparameters configuration.

TrainingMetrics

Training metrics collected during simulation.

TrainingSimulation

Simulated training scenario for reproducible ML experiments.

TrainingState

Training state captured at each epoch.

TrainingTrajectory

Training trajectory - sequence of training states.

Enums

TrainEvent

Simulated training event for journaling.

TrainingAnomaly

Training anomaly types for Jidoka detection.