https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo/Head https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo http://www.nanopub.org/nschema#hasAssertion https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo/assertion https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo http://www.nanopub.org/nschema#hasProvenance https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo/provenance https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo http://www.nanopub.org/nschema#hasPublicationInfo https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo/pubinfo https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://www.nanopub.org/nschema#Nanopublication https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo/assertion https://doi.org/10.48550/arXiv.2402.11163 http://purl.org/dc/terms/title KG-Agent: An Efficient Autonomous Agent Framework for Complex Reasoning over Knowledge Graph https://doi.org/10.48550/arXiv.2402.11163 http://purl.org/spar/cito/describes https://neverblink.eu/ontologies/llm-kg/methods#KgAgent https://doi.org/10.48550/arXiv.2402.11163 http://purl.org/spar/cito/discusses https://neverblink.eu/ontologies/llm-kg/methods#AutoGpt https://doi.org/10.48550/arXiv.2402.11163 http://purl.org/spar/cito/discusses https://neverblink.eu/ontologies/llm-kg/methods#BartSparql https://doi.org/10.48550/arXiv.2402.11163 http://purl.org/spar/cito/discusses https://neverblink.eu/ontologies/llm-kg/methods#ChatBd https://doi.org/10.48550/arXiv.2402.11163 http://purl.org/spar/cito/discusses https://neverblink.eu/ontologies/llm-kg/methods#EmbedKgqa https://doi.org/10.48550/arXiv.2402.11163 http://purl.org/spar/cito/discusses https://neverblink.eu/ontologies/llm-kg/methods#GraftNet https://doi.org/10.48550/arXiv.2402.11163 http://purl.org/spar/cito/discusses https://neverblink.eu/ontologies/llm-kg/methods#KbBinder https://doi.org/10.48550/arXiv.2402.11163 http://purl.org/spar/cito/discusses https://neverblink.eu/ontologies/llm-kg/methods#KvMemNet https://doi.org/10.48550/arXiv.2402.11163 http://purl.org/spar/cito/discusses https://neverblink.eu/ontologies/llm-kg/methods#MmReact https://doi.org/10.48550/arXiv.2402.11163 http://purl.org/spar/cito/discusses https://neverblink.eu/ontologies/llm-kg/methods#Nsm https://doi.org/10.48550/arXiv.2402.11163 http://purl.org/spar/cito/discusses https://neverblink.eu/ontologies/llm-kg/methods#Pangu https://doi.org/10.48550/arXiv.2402.11163 http://purl.org/spar/cito/discusses https://neverblink.eu/ontologies/llm-kg/methods#ProgPrompt https://doi.org/10.48550/arXiv.2402.11163 http://purl.org/spar/cito/discusses https://neverblink.eu/ontologies/llm-kg/methods#ReAct https://doi.org/10.48550/arXiv.2402.11163 http://purl.org/spar/cito/discusses https://neverblink.eu/ontologies/llm-kg/methods#Rgcn https://doi.org/10.48550/arXiv.2402.11163 http://purl.org/spar/cito/discusses https://neverblink.eu/ontologies/llm-kg/methods#RnnSparql https://doi.org/10.48550/arXiv.2402.11163 http://purl.org/spar/cito/discusses https://neverblink.eu/ontologies/llm-kg/methods#StructGpt https://doi.org/10.48550/arXiv.2402.11163 http://purl.org/spar/cito/discusses https://neverblink.eu/ontologies/llm-kg/methods#TransferNet https://doi.org/10.48550/arXiv.2402.11163 http://purl.org/spar/cito/discusses https://neverblink.eu/ontologies/llm-kg/methods#WebGpt https://doi.org/10.48550/arXiv.2402.11163 http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://www.w3.org/ns/prov#Entity https://neverblink.eu/ontologies/llm-kg/methods#AutoGpt http://purl.org/spar/fabio/hasURL https://github.com/Significant-Gravitas/AutoGPT https://neverblink.eu/ontologies/llm-kg/methods#AutoGpt http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://purl.org/spar/fabio/Workflow https://neverblink.eu/ontologies/llm-kg/methods#AutoGpt http://www.w3.org/2000/01/rdf-schema#comment AutoGPT is a representative LLM-based agent framework that empowers LLMs with long/short-term memory management and external tools. It is mentioned as a significant development in autonomously addressing user requests, providing context for the agent design. https://neverblink.eu/ontologies/llm-kg/methods#AutoGpt http://www.w3.org/2000/01/rdf-schema#label AutoGPT https://neverblink.eu/ontologies/llm-kg/methods#BartSparql http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://purl.org/spar/fabio/Workflow https://neverblink.eu/ontologies/llm-kg/methods#BartSparql http://www.w3.org/2000/01/rdf-schema#comment BART SPARQL is a method that uses the BART language model to generate SPARQL queries for KG-based question answering. It is included as a strong baseline in the experimental comparisons. https://neverblink.eu/ontologies/llm-kg/methods#BartSparql http://www.w3.org/2000/01/rdf-schema#label BART SPARQL https://neverblink.eu/ontologies/llm-kg/methods#ChatBd http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://purl.org/spar/fabio/Workflow https://neverblink.eu/ontologies/llm-kg/methods#ChatBd http://www.w3.org/2000/01/rdf-schema#comment ChatBD is an existing method for autonomous reasoning leveraging chain-of-thought and memory augmentation. The paper critically notes its reliance on strong closed-source LLM APIs and its inability to utilize external tools for specialized operations, contrasting with the KG-Agent's design. https://neverblink.eu/ontologies/llm-kg/methods#ChatBd http://www.w3.org/2000/01/rdf-schema#label ChatBD https://neverblink.eu/ontologies/llm-kg/methods#EmbedKgqa http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://purl.org/spar/fabio/Workflow https://neverblink.eu/ontologies/llm-kg/methods#EmbedKgqa http://www.w3.org/2000/01/rdf-schema#comment EmbedKGQA is a subgraph-based reasoning method, typically leveraging knowledge graph embeddings, that is used as a baseline for experimental comparison in the paper's evaluation of KG-Agent's performance. https://neverblink.eu/ontologies/llm-kg/methods#EmbedKgqa http://www.w3.org/2000/01/rdf-schema#label EmbedKGQA https://neverblink.eu/ontologies/llm-kg/methods#GraftNet http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://purl.org/spar/fabio/Workflow https://neverblink.eu/ontologies/llm-kg/methods#GraftNet http://www.w3.org/2000/01/rdf-schema#comment GraftNet is a baseline method for Knowledge Graph Question Answering. It is specifically mentioned and used for comparison in the evaluation of KG-Agent on the MetaQA dataset. https://neverblink.eu/ontologies/llm-kg/methods#GraftNet http://www.w3.org/2000/01/rdf-schema#label GraftNet https://neverblink.eu/ontologies/llm-kg/methods#KbBinder http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://purl.org/spar/fabio/Workflow https://neverblink.eu/ontologies/llm-kg/methods#KbBinder http://www.w3.org/2000/01/rdf-schema#comment KB-BINDER is discussed as a prior method for LLM-KG reasoning that employs a pre-defined interaction workflow between the LLM and KG, which the authors contrast with their proposed KG-Agent's autonomous approach. https://neverblink.eu/ontologies/llm-kg/methods#KbBinder http://www.w3.org/2000/01/rdf-schema#label KB-BINDER https://neverblink.eu/ontologies/llm-kg/methods#KgAgent http://purl.org/dc/terms/subject https://neverblink.eu/ontologies/llm-kg/categories#SynergizedReasoning https://neverblink.eu/ontologies/llm-kg/methods#KgAgent http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://purl.org/spar/fabio/Workflow https://neverblink.eu/ontologies/llm-kg/methods#KgAgent http://www.w3.org/2000/01/rdf-schema#comment KG-Agent is an autonomous agent framework that integrates an instruction-tuned LLM, a multifunctional toolbox, a KG-based executor, and knowledge memory. It synergizes the LLM's planning and decision-making capabilities with KG's structured knowledge and operations through an iterative process of tool selection and memory updates, explicitly treating the LLM as an agent interacting with the KG for complex reasoning to enhance LLM performance. https://neverblink.eu/ontologies/llm-kg/methods#KgAgent http://www.w3.org/2000/01/rdf-schema#label KG-Agent https://neverblink.eu/ontologies/llm-kg/methods#KgAgent https://neverblink.eu/ontologies/llm-kg/hasTopCategory https://neverblink.eu/ontologies/llm-kg/top-categories#SynergizedLLMKG https://neverblink.eu/ontologies/llm-kg/methods#KvMemNet http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://purl.org/spar/fabio/Workflow https://neverblink.eu/ontologies/llm-kg/methods#KvMemNet http://www.w3.org/2000/01/rdf-schema#comment KVMemNet is a subgraph-based reasoning method used as a baseline in the experimental evaluations to compare the performance of KG-Agent against established approaches in KG-based question answering. https://neverblink.eu/ontologies/llm-kg/methods#KvMemNet http://www.w3.org/2000/01/rdf-schema#label KVMemNet https://neverblink.eu/ontologies/llm-kg/methods#MmReact http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://purl.org/spar/fabio/Workflow https://neverblink.eu/ontologies/llm-kg/methods#MmReact http://www.w3.org/2000/01/rdf-schema#comment MM-REACT is introduced as an LLM-based agent framework tailored for multi-modal scenarios. It serves as an example of specialized agents in the related work section, showcasing the diversity of agent applications. https://neverblink.eu/ontologies/llm-kg/methods#MmReact http://www.w3.org/2000/01/rdf-schema#label MM-REACT https://neverblink.eu/ontologies/llm-kg/methods#Nsm http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://purl.org/spar/fabio/Workflow https://neverblink.eu/ontologies/llm-kg/methods#Nsm http://www.w3.org/2000/01/rdf-schema#comment NSM (Neural Symbolic Machine) is a baseline method for KGQA. It is included in the experimental comparisons, particularly for the MetaQA dataset, to contextualize the performance of KG-Agent. https://neverblink.eu/ontologies/llm-kg/methods#Nsm http://www.w3.org/2000/01/rdf-schema#label NSM https://neverblink.eu/ontologies/llm-kg/methods#Pangu http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://purl.org/spar/fabio/Workflow https://neverblink.eu/ontologies/llm-kg/methods#Pangu http://www.w3.org/2000/01/rdf-schema#comment Pangu is presented as another existing method for LLM-KG reasoning characterized by a pre-defined interaction mechanism. It serves as a comparative baseline to highlight the flexible and autonomous nature of the KG-Agent framework. https://neverblink.eu/ontologies/llm-kg/methods#Pangu http://www.w3.org/2000/01/rdf-schema#label Pangu https://neverblink.eu/ontologies/llm-kg/methods#ProgPrompt http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://purl.org/spar/fabio/Workflow https://neverblink.eu/ontologies/llm-kg/methods#ProgPrompt http://www.w3.org/2000/01/rdf-schema#comment ProgPrompt is mentioned as an LLM-based agent framework designed for real-life environments. It is included in the discussion of domain-specific agents, illustrating how agents are adapted to particular contexts. https://neverblink.eu/ontologies/llm-kg/methods#ProgPrompt http://www.w3.org/2000/01/rdf-schema#label ProgPrompt https://neverblink.eu/ontologies/llm-kg/methods#ReAct http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://purl.org/spar/fabio/Workflow https://neverblink.eu/ontologies/llm-kg/methods#ReAct http://www.w3.org/2000/01/rdf-schema#comment ReAct is a general LLM-based agent framework that converts LLMs into language agents capable of interacting with external environments, receiving feedback, and generating actions. It is discussed as foundational work in the broader field of LLM agents. https://neverblink.eu/ontologies/llm-kg/methods#ReAct http://www.w3.org/2000/01/rdf-schema#label ReAct https://neverblink.eu/ontologies/llm-kg/methods#Rgcn http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://purl.org/spar/fabio/Workflow https://neverblink.eu/ontologies/llm-kg/methods#Rgcn http://www.w3.org/2000/01/rdf-schema#comment RGCN (Relational Graph Convolutional Networks) is a graph neural network-based method for KG reasoning. It is included as a baseline in the experimental section to benchmark the performance of the proposed KG-Agent. https://neverblink.eu/ontologies/llm-kg/methods#Rgcn http://www.w3.org/2000/01/rdf-schema#label RGCN https://neverblink.eu/ontologies/llm-kg/methods#RnnSparql http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://purl.org/spar/fabio/Workflow https://neverblink.eu/ontologies/llm-kg/methods#RnnSparql http://www.w3.org/2000/01/rdf-schema#comment RNN SPARQL is a method that employs Recurrent Neural Networks to generate SPARQL queries for KG-based question answering. It is utilized as a baseline method in the experimental setup to evaluate KG-Agent's effectiveness. https://neverblink.eu/ontologies/llm-kg/methods#RnnSparql http://www.w3.org/2000/01/rdf-schema#label RNN SPARQL https://neverblink.eu/ontologies/llm-kg/methods#StructGpt http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://purl.org/spar/fabio/Workflow https://neverblink.eu/ontologies/llm-kg/methods#StructGpt http://www.w3.org/2000/01/rdf-schema#comment Struct-GPT is discussed as a method that utilizes a pre-defined interaction strategy between LLMs and KGs for reasoning. It is used as a baseline for comparison, particularly when evaluating the transferability of methods to domain-specific KGs. https://neverblink.eu/ontologies/llm-kg/methods#StructGpt http://www.w3.org/2000/01/rdf-schema#label Struct-GPT https://neverblink.eu/ontologies/llm-kg/methods#TransferNet http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://purl.org/spar/fabio/Workflow https://neverblink.eu/ontologies/llm-kg/methods#TransferNet http://www.w3.org/2000/01/rdf-schema#comment TransferNet is a baseline method for Knowledge Graph Question Answering. It is used in the experimental evaluation, specifically compared against KG-Agent on the MetaQA dataset, particularly for its supervised fine-tuning performance. https://neverblink.eu/ontologies/llm-kg/methods#TransferNet http://www.w3.org/2000/01/rdf-schema#label TransferNet https://neverblink.eu/ontologies/llm-kg/methods#WebGpt http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://purl.org/spar/fabio/Workflow https://neverblink.eu/ontologies/llm-kg/methods#WebGpt http://www.w3.org/2000/01/rdf-schema#comment WebGPT is highlighted as an LLM-based agent specifically designed for web-browsing environments. It is cited as an example of domain-specific agent frameworks in the related work section. https://neverblink.eu/ontologies/llm-kg/methods#WebGpt http://www.w3.org/2000/01/rdf-schema#label WebGPT https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo/provenance https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo/assertion http://www.w3.org/ns/prov#wasAttributedTo https://neverblink.eu/ontologies/llm-kg/agent https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo/assertion http://www.w3.org/ns/prov#wasDerivedFrom https://doi.org/10.48550/arXiv.2402.11163 https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo/pubinfo https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo http://purl.org/dc/terms/created 2026-03-13T16:04:19.154Z https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo http://purl.org/dc/terms/creator https://neverblink.eu/ontologies/llm-kg/agent https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo http://purl.org/nanopub/x/hasNanopubType https://neverblink.eu/ontologies/llm-kg/PaperAssessmentResult https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo http://purl.org/nanopub/x/supersedes https://w3id.org/np/RAulYmIvuoBWivv_3hbcSjw1QWcq8tkaXnNJcVruVktqs https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo http://www.w3.org/2000/01/rdf-schema#label LLM-KG assessment for paper 10.48550/arXiv.2402.11163 https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo/sig http://purl.org/nanopub/x/hasAlgorithm RSA https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo/sig http://purl.org/nanopub/x/hasPublicKey MIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEAwNz2QK3SEifno78S7+48zUB0xpTex3mAzW73ZimHqNcdEMU5/apslrGrTHGFAt/Chocgo++r6JQp5ygY7NyJHGWdaIqnt85pjX4PbNfLAvapyUO00qZP34fY61w4eZ9UMtleWEsmZKRtQPyJ8ODl46i/rfPuZlcJGpM9Nmy5mpGWuepqIEvF4a/t7pLVeCEDFSYXT+yaiygt6ynIK5f7TtEDhZpeUf/Q74WhMPJXm4yTU/hqOX4IW+50kWHNArGGZwUaXwzyG6M3Zd6UMModryGkLqS4H/MSE3ZA1Ylnms7BfWLEXhMWlaKi6HRV4nGRDLhxVSi9LSRi3LWKLhNIIQIDAQAB https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo/sig http://purl.org/nanopub/x/hasSignature BIyn6+rueJUGaNC3Fjg82/u6nCscV/bNJ9BdqYtQ5xOub0iBfY4nkZ/HRqlZxfRRDt8rVUzIZ+OcGkDGZove3kaZvuwP6VKSdEL468ueRSPCWDVX2NtAGpLzVWFggro23hFqxCjSQ+BB0N5cWMQQQJUGM9Tsw4yqWktqCfYh4oqxIfJ/Yb6sQ+WRizCll6+sv1H76ZWMpIrkQdSGSNqbWxpLrTLUb7CXLVar/LXs2i2OV5tJCNpKEAE4xVhUa8tq2TDTt/ryTMTvt4sdKOEsXq0l9LlDPHWIGIznHoGDKg9JtijihPaji69h3g498CVwhRkLtzYYXNd+l/eo8YYHwg== https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo/sig http://purl.org/nanopub/x/hasSignatureTarget https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo https://w3id.org/np/RAEHejEFl7_rfCADkt9USXyxjhbPQ-TJvsZOMKqcoiGxo/sig http://purl.org/nanopub/x/signedBy https://neverblink.eu/ontologies/llm-kg/agent