Module models 

Embedding model implementations

This module provides various knowledge graph embedding models including:

• TransE: Translation-based embeddings
• ComplEx: Complex number embeddings for asymmetric relations
• DistMult: Bilinear diagonal model
• RotatE: Rotation-based embeddings
• HolE: Holographic embeddings using circular correlation
• ConvE: Convolutional embeddings with 2D CNNs (optional)
• TuckER: Tucker decomposition based embeddings (optional)
• TransformerEmbedding: Transformer-based embeddings (BERT, RoBERTa, etc.)
• GNNEmbedding: Graph Neural Network embeddings (GCN, GraphSAGE, GAT, etc.)
• OntologyAwareEmbedding: Embeddings that respect RDF/OWL ontology constraints

Re-exports

pub use complex::ComplEx;

pub use distmult::DistMult;

pub use gnn::AggregationType;

pub use gnn::GNNConfig;

pub use gnn::GNNEmbedding;

pub use gnn::GNNType;

pub use hole::HoLE;

pub use hole::HoLEConfig;

pub use ontology::OntologyAwareConfig;

pub use ontology::OntologyAwareEmbedding;

pub use ontology::OntologyConstraints;

pub use ontology::OntologyRelation;

pub use rotate::RotatE;

pub use transe::TransE;

pub use transformer::PoolingStrategy;

pub use transformer::TransformerConfig;

pub use transformer::TransformerEmbedding;

pub use transformer::TransformerType;

pub use conve::ConvE;

pub use conve::ConvEConfig;

pub use tucker::TuckER;

pub use quatd::QuatD;

pub use advanced_models::Lcg as AdvancedLcg;

pub use advanced_models::PairRE;

pub use advanced_models::Rescal;

pub use advanced_models::RotatEPlus;

pub use scirs_neural::ActivationType;

pub use scirs_neural::OptimizerType;

pub use scirs_neural::SciRS2NeuralConfig;

pub use scirs_neural::SciRS2NeuralEmbedding;

pub use graphsage::AggregatorType;

pub use graphsage::GraphData;

pub use graphsage::GraphSage;

pub use graphsage::GraphSageConfig;

pub use graphsage::GraphSageEmbeddings;

pub use graphsage::GraphSageTrainingMetrics;

pub use graphsage::SimpleLcg;

pub use gat_basic::Gat;

pub use gat_basic::GatConfig;

pub use gat_basic::GatEmbeddings;

pub use kg_embeddings::deserialize_embeddings;

pub use kg_embeddings::serialize_embeddings;

pub use kg_embeddings::DistMult as KgDistMult;

pub use kg_embeddings::KgEmbeddingConfig;

pub use kg_embeddings::KgEmbeddings;

pub use kg_embeddings::KgError;

pub use kg_embeddings::KgModel;

pub use kg_embeddings::KgResult;

pub use kg_embeddings::KgTriple;

pub use kg_embeddings::LinkPredictionEvaluator;

pub use kg_embeddings::RotatE as KgRotatE;

pub use kg_embeddings::TrainingHistory;

pub use kg_embeddings::TransE as KgTransE;

pub use base::*;

pub use common::*;

Modules

advanced_models

Advanced Knowledge Graph Embedding Models

base

Base functionality shared across embedding models

common

Common utilities and functions used across embedding models

complex

ComplEx: Complex Embeddings for Simple Link Prediction

conve

ConvE (Convolutional Embeddings) Model

distmult

DistMult: Embedding Entities and Relations for Learning and Inference in Knowledge Bases

gat_basic

Graph Attention Networks (GAT) - Basic Implementation

gnn

Graph Neural Network (GNN) embedding models

graphsage

GraphSAGE: Inductive Representation Learning on Large Graphs

hole

HolE (Holographic Embeddings) Model

kg_embeddings

Knowledge Graph Embedding algorithms for link prediction.

ontology

Ontology-aware embedding models

quatd

QuatE: Quaternion Embeddings for Knowledge Graph Completion

rotate

RotatE: Rotation-based Knowledge Graph Embeddings

scirs_neural

SciRS2 Neural Network Integration for Enhanced Embeddings

transe

TransE: Translating Embeddings for Modeling Multi-relational Data

transformer

Transformer-based embedding models module

tucker

TuckER: Tucker Decomposition for Knowledge Graph Embeddings