Struct LearningEngine 

pub struct LearningEngine { /* private fields */ }

Learning engine that improves search quality over time.

The engine uses several mechanisms inspired by the ruvector-gnn crate:

• Experience replay for stable learning
• GNN-style message passing for relevance propagation
• Elastic weight consolidation to prevent catastrophic forgetting

Implementations

impl LearningEngine

pub fn new(dimensions: usize, learning_rate: f32) -> Self

Create a new learning engine.

pub fn rerank( &self, query_embedding: &[f32], candidates: Vec<(Uuid, f32)>, vectors: &DashMap<Uuid, Vec<f32>>, ) -> Vec<(Uuid, f32)>

Re-rank candidates based on learned patterns.

pub fn record_query( &mut self, query_embedding: &[f32], results: &[SearchResult], )

Record a query and its results for learning.

pub fn record_feedback(&mut self, result_embedding: &[f32], positive: bool)

Record user feedback on a search result.

pub fn query_count(&self) -> u64

Get the current query count.

pub fn stats(&self) -> LearningStats

Get learning statistics.