HighlightCurated by Aramai Editorialopenalex.org

TRACE the Evidence: Constructing Knowledge-Grounded Reasoning Chains for Retrieval-Augmented Generation

A paper proposing TRACE, a method for constructing knowledge-grounded reasoning chains to enhance multi-hop question answering.

The authors propose TRACE, a method that constructs knowledge-grounded reasoning chains to improve multi-hop question answering. TRACE uses a KG Generator and Autoregressive Reasoning Chain Constructor to build reasoning chains from retrieved documents. Experimental results show an average performance improvement of up to 14.03% compared to using all retrieved documents.

Based on: TRACE the Evidence: Constructing Knowledge-Grounded Reasoning Chains for Retrieval-Augmented Generation

HighlightCurated by Aramai Editorialopenalex.org

Mindful-RAG: A Study of Points of Failure in Retrieval Augmented Generation

A study on the limitations and failures of knowledge graph-based retrieval-augmented generation systems.

The paper identifies eight key areas of concern in existing KG-based RAG methods, including misinterpretation of question context and incorrect relation mapping. It proposes a new approach, Mindful-RAG, which re-engineers the retrieval process to be more intent-driven and contextually aware. The authors aim to improve the reliability and effectiveness of KG-RAG systems through enhanced reasoning capabilities and structural limitations of knowledge graphs.

Based on: Mindful-RAG: A Study of Points of Failure in Retrieval Augmented Generation

HighlightCurated by Aramai EditorialarXiv (Cornell University)

Retrieval-Augmented Generation with Graphs (GraphRAG)

A survey on retrieval-augmented generation with graphs, a technique for enhancing downstream tasks by retrieving information from external sources.

The paper presents a comprehensive survey of GraphRAG, a framework that combines graph-structured data with retrieval-augmented generation. It defines key components and reviews techniques tailored to different domains. The authors also discuss research challenges and potential directions for future work.

Based on: Retrieval-Augmented Generation with Graphs (GraphRAG) · arXiv (Cornell University)

HighlightCurated by Aramai EditorialarXiv (Cornell University)

G-Retriever: Retrieval-Augmented Generation for Textual Graph Understanding and Question Answering

A method for question answering on textual graphs using retrieval-augmented generation.

The authors propose G-Retriever, a framework for question-answering on textual graphs. They introduce a new approach called retrieval-augmented generation (RAG) and formulate the task as a Prize-Collecting Steiner Tree optimization problem to mitigate hallucination. The method is evaluated on various textual graph tasks and outperforms baselines.

Based on: G-Retriever: Retrieval-Augmented Generation for Textual Graph Understanding and Question Answering · arXiv (Cornell University)

HighlightCurated by Aramai Editorialopenalex.org

Enhancing Knowledge Graph Completion with Retrieval-Augmented Generation Using Large Language Models

A study introducing a framework for Knowledge Graph Completion using Large Language Models and Retrieval-Augmented Generation.

The authors propose an innovative framework for Knowledge Graph Completion leveraging Large Language Models. The framework treats KG triples as textual prompts to retrieve relevant information from knowledge bases, generating contextually accurate responses. A retrieval reranking strategy refines predictions by incorporating outputs from a pre-trained KGC model.

Based on: Enhancing Knowledge Graph Completion with Retrieval-Augmented Generation Using Large Language Models

HighlightCurated by Aramai Editorialopenalex.org

Patent Response System Optimised for Faithfulness: Procedural Knowledge Embodiment with Knowledge Graph and Retrieval Augmented Generation

A proposed system for generating faithful and unbiased patent responses using a knowledge graph and retrieval augmented generation.

The authors propose the Patent Response System Optimised for Faithfulness (PRO), which incorporates procedural knowledge and uses a tailored large language model to generate patent responses. PRO outperforms GPT-4 in terms of faithfulness, reducing unfaithfulness across six error types. The system's effectiveness is demonstrated through experimental results.

Based on: Patent Response System Optimised for Faithfulness: Procedural Knowledge Embodiment with Knowledge Graph and Retrieval Augmented Generation

HighlightCurated by Aramai EditorialElectronics

CRP-RAG: A Retrieval-Augmented Generation Framework for Supporting Complex Logical Reasoning and Knowledge Planning

A framework that enhances Large Language Models by retrieving relevant knowledge.

The CRP-RAG framework addresses limitations in existing Retrieval-Augmented Generation methods. It employs reasoning graphs to model complex query reasoning processes and guides knowledge retrieval, aggregation, and evaluation through these graphs. This approach outperforms baseline models in open-domain QA, multi-hop reasoning, and factual verification.

Based on: CRP-RAG: A Retrieval-Augmented Generation Framework for Supporting Complex Logical Reasoning and Knowledge Planning · Electronics

HighlightCurated by Aramai Editorialopenalex.org

Graph Retrieval-Augmented Generation for Large Language Models: A Survey

A survey on incorporating Knowledge Graphs with Large Language Model Retrieval-Augmented Generation.

The paper surveys work that combines Knowledge Graphs with Large Language Model Retrieval-Augmented Generation to optimize model performance. It highlights the importance of precise document selection and noise-free corpora for expert tasks. The authors aim to provide a comprehensive understanding of this research area for future work.

Based on: Graph Retrieval-Augmented Generation for Large Language Models: A Survey

HighlightCurated by Aramai Editorialopenalex.org

Scalable Extraction and Adoption of Shapes for Improving Data Quality and Query Processing in Knowledge Graphs

A thesis proposing techniques to improve data quality, efficient data access, and interoperability in Knowledge Graphs.

The resource proposes Quality Shapes Extraction (QSE) and SHACTOR to enhance data quality in Knowledge Graphs. It also introduces 'shapes statistics' for optimizing SPARQL query processing over KGs. The approach is demonstrated on both synthetic and real-world datasets, showing potential improvements in query performance.

Based on: Scalable Extraction and Adoption of Shapes for Improving Data Quality and Query Processing in Knowledge Graphs

HighlightCurated by Aramai EditorialPeerJ Computer Science

Evaluating FAIR Digital Object and Linked Data as distributed object systems

A paper evaluating the FAIR Digital Object concept and its implementations.

The authors evaluate FAIR Digital Object (FDO) as a global distributed object system using five conceptual frameworks. They compare FDO with established Linked Data practices and Web architecture, providing recommendations for both communities. The paper discusses the history of the Semantic Web and its relevance to FDO adoption.

Based on: Evaluating FAIR Digital Object and Linked Data as distributed object systems · PeerJ Computer Science

HighlightCurated by Aramai Editorialopenalex.org

Retrieval-Augmented Generation with Knowledge Graphs for Customer Service Question Answering

A novel customer service question-answering method that amalgamates RAG with a knowledge graph.

The paper introduces a method that constructs a knowledge graph from historical issues to improve retrieval accuracy and answering quality. It combines retrieval-augmented generation (RAG) with a knowledge graph, preserving intra-issue structure and inter-issue relations. Empirical assessments show improved performance over baseline methods in key metrics.

Based on: Retrieval-Augmented Generation with Knowledge Graphs for Customer Service Question Answering

HighlightCurated by Aramai EditorialInformation

Construction of Knowledge Graphs: Current State and Challenges

A research paper on the current state and challenges of constructing knowledge graphs.

The authors discuss the main graph models for knowledge graphs, introduce requirements for future construction pipelines, and evaluate the state-of-the-art. They identify areas in need of further research and improvement. The paper provides an overview of necessary steps to build high-quality knowledge graphs, including metadata management and quality assurance.

Based on: Construction of Knowledge Graphs: Current State and Challenges · Information