Struct FeatureFusion 

pub struct FeatureFusion { /* private fields */ }

Feature fusion module

Combines GNN embeddings, min-cut features, and boundary features into a unified representation for error prediction.

Implementations

impl FeatureFusion

pub fn new(config: FusionConfig) -> Result<Self>

Create a new feature fusion module

pub fn fuse_simple( &self, gnn_features: &Array2<f32>, mincut_features: &Array2<f32>, ) -> Result<Array2<f32>>

Simple fuse for GNN and MinCut features only

Arguments

• gnn_features - GNN embeddings (num_nodes, gnn_dim)
• mincut_features - MinCut features (num_nodes, mincut_dim)

Returns

Fused features (num_nodes, output_dim)

pub fn fuse( &self, gnn_features: &Array2<f32>, mincut_features: &Array2<f32>, boundary_features: &Array2<f32>, confidences: Option<&[f32]>, ) -> Result<Array2<f32>>

Fuse features from multiple sources

Arguments

• gnn_features - GNN embeddings (num_nodes, gnn_dim)
• mincut_features - MinCut features (num_nodes, mincut_dim)
• boundary_features - Boundary features (num_nodes, 8)
• confidences - Optional confidence scores for adaptive weighting

Returns

Fused features (num_nodes, output_dim)

pub fn fuse_all( &self, gnn_embeddings: &Array2<f32>, adjacency: &HashMap<usize, Vec<usize>>, edge_weights: &HashMap<(usize, usize), f32>, positions: &[(f32, f32)], grid_size: usize, ) -> Result<Array2<f32>>

Convenience method to compute all features and fuse them

Arguments

• gnn_embeddings - GNN node embeddings
• adjacency - Graph adjacency list
• edge_weights - Edge weights for min-cut
• positions - Node positions
• grid_size - Grid size for boundary computation

pub fn config(&self) -> &FusionConfig

Get configuration

pub fn output_dim(&self) -> usize

Get output dimension

Trait Implementations

impl Clone for FeatureFusion

fn clone(&self) -> FeatureFusion

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for FeatureFusion

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for FeatureFusion

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for FeatureFusion

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.