sub:assertion {
<
https://doi.org/10.48550/arXiv.2506.21556>
dct:title "VAT-KG: KNOWLEDGE-INTENSIVE MULTIMODAL KNOWLEDGE GRAPH DATASET FOR RETRIEVAL-AUGMENTED GENERATION" ;
<
http://purl.org/spar/cito/describes> <
https://neverblink.eu/ontologies/llm-kg/methods#multimodalRagFramework> , <
https://neverblink.eu/ontologies/llm-kg/methods#vatkgConstructionPipeline> ;
<
http://purl.org/spar/cito/discusses> <
https://neverblink.eu/ontologies/llm-kg/methods#gemini2Dot5> , <
https://neverblink.eu/ontologies/llm-kg/methods#gpt4o> , <
https://neverblink.eu/ontologies/llm-kg/methods#m2ConceptBase> , <
https://neverblink.eu/ontologies/llm-kg/methods#qwen2Dot5Omni> , <
https://neverblink.eu/ontologies/llm-kg/methods#videoLlama2> , <
https://neverblink.eu/ontologies/llm-kg/methods#wikidata5M> ;
a prov:Entity .
<
https://neverblink.eu/ontologies/llm-kg/methods#gemini2Dot5>
a <
http://purl.org/spar/fabio/Workflow> ;
rdfs:label "Gemini-2.5" .
<
https://neverblink.eu/ontologies/llm-kg/methods#gpt4o>
a <
http://purl.org/spar/fabio/Workflow> ;
rdfs:label "GPT-4o" .
<
https://neverblink.eu/ontologies/llm-kg/methods#m2ConceptBase>
a <
http://purl.org/spar/fabio/Workflow> ;
rdfs:label "M2ConceptBase" .
<
https://neverblink.eu/ontologies/llm-kg/methods#multimodalRagFramework>
dct:subject <
https://neverblink.eu/ontologies/llm-kg/categories#SynergizedReasoning> ;
a <
http://purl.org/spar/fabio/Workflow> ;
rdfs:comment "This framework integrates VAT-KG with Multimodal Large Language Models (MLLMs) for question answering across diverse modalities. It synergistically combines retrieval from VAT-KG with MLLM generation and includes a \"Retrieval Checker\" that uses the MLLM's text encoder to filter retrieved KG triplets, indicating a unified reasoning process where both LLM and KG representations are actively combined." ;
rdfs:label "Multimodal RAG Framework" ;
<
https://neverblink.eu/ontologies/llm-kg/hasTopCategory> <
https://neverblink.eu/ontologies/llm-kg/top-categories#SynergizedLLMKG> .
<
https://neverblink.eu/ontologies/llm-kg/methods#qwen2Dot5Omni>
a <
http://purl.org/spar/fabio/Workflow> ;
rdfs:label "Qwen2.5-Omni" .
<
https://neverblink.eu/ontologies/llm-kg/methods#vatkgConstructionPipeline>
dct:subject <
https://neverblink.eu/ontologies/llm-kg/categories#LLMAugmentedKGConstruction> ;
a <
http://purl.org/spar/fabio/Workflow> ;
rdfs:comment "This method proposes a four-step pipeline for constructing the Visual-Audio-Text Knowledge Graph (VAT-KG). It explicitly leverages LLMs for \"Knowledge-Intensive Recaptioning\" to enrich textual data and \"Multimodal Triplet Grounding\" to extract knowledge-intensive triplets, directly supporting the construction of the knowledge graph. Thus, LLMs are used to improve a KG task, specifically KG construction." ;
rdfs:label "VAT-KG Construction Pipeline" ;
<
https://neverblink.eu/ontologies/llm-kg/hasTopCategory> <
https://neverblink.eu/ontologies/llm-kg/top-categories#LLMAugmentedKG> .
<
https://neverblink.eu/ontologies/llm-kg/methods#videoLlama2>
a <
http://purl.org/spar/fabio/Workflow> ;
rdfs:label "Video-LLaMA2" .
<
https://neverblink.eu/ontologies/llm-kg/methods#wikidata5M>
a <
http://purl.org/spar/fabio/Workflow> ;
rdfs:label "Wikidata5M" .
}