sub:assertion {
<
https://doi.org/10.48550/arXiv.2403.05881>
dct:title "KG-Rank: Enhancing Large Language Models for Medical QA with Knowledge Graphs and Ranking Techniques" ;
<
http://purl.org/spar/cito/describes> <
https://neverblink.eu/ontologies/llm-kg/methods#KgRank> ;
<
http://purl.org/spar/cito/discusses> <
https://neverblink.eu/ontologies/llm-kg/methods#Almanac> , <
https://neverblink.eu/ontologies/llm-kg/methods#ChatEnt> ;
a prov:Entity .
<
https://neverblink.eu/ontologies/llm-kg/methods#Almanac>
a <
http://purl.org/spar/fabio/Workflow> ;
rdfs:comment "Almanac is discussed as previous research that leverages external medical knowledge to enhance the accuracy and reliability of LLM-generated content. It is mentioned in the introduction as a related work that attempted to address the challenge of LLM factual inconsistency, providing context for the problem KG-Rank aims to solve." ;
rdfs:label "Almanac" .
<
https://neverblink.eu/ontologies/llm-kg/methods#ChatEnt>
a <
http://purl.org/spar/fabio/Workflow> ;
rdfs:comment "ChatENT is discussed as previous research that leverages external medical knowledge to enhance the accuracy and reliability of LLM-generated content. Similar to Almanac, it is cited in the introduction as related work demonstrating the use of external knowledge to improve LLMs, setting the stage for KG-Rank's novel approach to knowledge integration." ;
rdfs:label "ChatENT" .
<
https://neverblink.eu/ontologies/llm-kg/methods#KgRank>
dct:subject <
https://neverblink.eu/ontologies/llm-kg/categories#KGEnhancedLLMInference> ;
a <
http://purl.org/spar/fabio/Workflow> ;
rdfs:comment "KG-Rank is an augmented LLM framework that integrates a medical Knowledge Graph (KG) with multiple ranking and re-ranking techniques to improve the factual consistency of long-form question answering in the medical domain. It works by identifying medical entities in a question, retrieving related KG triples, and then applying techniques like Similarity Ranking, Answer Expansion Ranking, and Maximal Marginal Relevance Ranking, followed by re-ranking using models like MedCPT, to refine the information provided to the LLM during inference for answer generation." ;
rdfs:label "KG-Rank" ;
<
https://neverblink.eu/ontologies/llm-kg/hasTopCategory> <
https://neverblink.eu/ontologies/llm-kg/top-categories#KGEnhancedLLM> .
}