sub:assertion {
  <https://doi.org/10.48550/arXiv.2403.07311> dct:title "Knowledge Graph Large Language Model (KG-LLM) for Link Prediction" ;
    <http://purl.org/spar/cito/describes> <https://neverblink.eu/ontologies/llm-kg/methods#KnowledgeGraphLargeLanguageModel> ;
    <http://purl.org/spar/cito/discusses> <https://neverblink.eu/ontologies/llm-kg/methods#Analogy> , <https://neverblink.eu/ontologies/llm-kg/methods#CompleX> , <https://neverblink.eu/ontologies/llm-kg/methods#ConGLR> , <https://neverblink.eu/ontologies/llm-kg/methods#ConvRot> , <https://neverblink.eu/ontologies/llm-kg/methods#DistMult> , <https://neverblink.eu/ontologies/llm-kg/methods#RESCAL> , <https://neverblink.eu/ontologies/llm-kg/methods#TransE> , <https://neverblink.eu/ontologies/llm-kg/methods#WsGAT> ;
    a prov:Entity .
  <https://neverblink.eu/ontologies/llm-kg/methods#Analogy> a <http://purl.org/spar/fabio/Workflow> ;
    rdfs:comment "Analogy is a knowledge graph embedding model designed to capture graph structures. It is used as a baseline for comparison against the KG-LLM framework in multi-hop link prediction tasks." ;
    rdfs:label "Analogy" .
  <https://neverblink.eu/ontologies/llm-kg/methods#CompleX> a <http://purl.org/spar/fabio/Workflow> ;
    rdfs:comment "CompleX utilizes complex embeddings for entities and relations to capture asymmetric relationships in knowledge graphs. It is used as a baseline for comparison against the KG-LLM framework in multi-hop link prediction tasks." ;
    rdfs:label "CompleX" .
  <https://neverblink.eu/ontologies/llm-kg/methods#ConGLR> a <http://purl.org/spar/fabio/Workflow> ;
    rdfs:comment "ConGLR is a GNN-based model that leverages context-aware graph representation learning and logical reasoning for improved inductive relation prediction. It is used as a baseline for comparison against the KG-LLM framework." ;
    rdfs:label "ConGLR" .
  <https://neverblink.eu/ontologies/llm-kg/methods#ConvRot> a <http://purl.org/spar/fabio/Workflow> ;
    rdfs:comment "ConvRot integrates convolutional networks and rotational embeddings to enhance link prediction performance in knowledge graphs. It is used as a GNN-based baseline for comparison against the KG-LLM framework." ;
    rdfs:label "ConvRot" .
  <https://neverblink.eu/ontologies/llm-kg/methods#DistMult> a <http://purl.org/spar/fabio/Workflow> ;
    rdfs:comment "DistMult represents relations as diagonal matrices for simplicity and efficiency in knowledge graph embedding. It is used as a baseline for comparison against the KG-LLM framework in multi-hop link prediction tasks." ;
    rdfs:label "DistMult" .
  <https://neverblink.eu/ontologies/llm-kg/methods#KnowledgeGraphLargeLanguageModel> dct:subject <https://neverblink.eu/ontologies/llm-kg/categories#LLMAugmentedKGCompletion> ;
    a <http://purl.org/spar/fabio/Workflow> ;
    rdfs:comment "The KG-LLM framework proposes a novel approach that converts multi-hop knowledge graph paths into structured natural language Chain-of-Thought prompts. These prompts are then used to instruction fine-tune large language models, incorporating In-Context Learning to enhance their performance in multi-hop link prediction and multi-hop relation prediction tasks, leveraging LLMs to complete missing facts within KGs." ;
    rdfs:label "Knowledge Graph Large Language Model (KG-LLM)" ;
    <https://neverblink.eu/ontologies/llm-kg/hasTopCategory> <https://neverblink.eu/ontologies/llm-kg/top-categories#LLMAugmentedKG> .
  <https://neverblink.eu/ontologies/llm-kg/methods#RESCAL> a <http://purl.org/spar/fabio/Workflow> ;
    rdfs:comment "RESCAL is a tensor factorization method used for knowledge graph embedding that captures rich interactions between entities and relations. It is used as a baseline for comparison against the KG-LLM framework in multi-hop link prediction tasks." ;
    rdfs:label "RESCAL" .
  <https://neverblink.eu/ontologies/llm-kg/methods#TransE> a <http://purl.org/spar/fabio/Workflow> ;
    rdfs:comment "TransE is a traditional knowledge graph embedding model that represents relationships as translations in the embedding space. It is used as a baseline for comparison against the KG-LLM framework in multi-hop link prediction tasks." ;
    rdfs:label "TransE" .
  <https://neverblink.eu/ontologies/llm-kg/methods#WsGAT> a <http://purl.org/spar/fabio/Workflow> ;
    rdfs:comment "wsGAT is a graph attention network model that uses weighted self-attention mechanisms to perform various knowledge graph tasks, including link prediction. It is used as a GNN-based baseline for comparison against the KG-LLM framework." ;
    rdfs:label "wsGAT" .
}