RAGE-KG
Retrieval-Augmented Generation Enabled by Knowledge Graphs

Overview

Retrieval-Augmented Generation Enabled by Knowledge Graphs (RAGE-KG) aims to explore the state of the art and go beyond in integrating Retrieval-Augmented Generation (RAG) with the Semantic Web infrastructure as well as the synergies between Large Language Models (LLMs) and the Linked Open Data (LOD) ecosystem. The workshop seeks to foster innovative RAG architectures relying on Semantic Web standards and new approaches to make LOD usable by LLMs, enhancing their ability to generate reliable, verifiable and context-aware responses based on structured, decentralized and authoritative data sources. The workshop is majorly focused on techniques for fine-tuning, prompting and otherwise integrating LLMs with KGs. Additionally, it will examine the development and application of RAG architectures across various use cases.

The workshop will feature two tracks:

• Novel Work Track: Invites full research papers (7-12 pages), short research papers (3-6 pages), position papers (6-8 pages), resource papers (8-12 pages), and demo papers (6-8 pages). Selected papers will be published in the workshop proceedings.
• Previously Published Work Track: Invites previously published full papers, resource papers, and demo papers. Papers will not be published again but provide an opportunity to discuss relevant work with the community.

Papers should be formatted according to CEUR Publications format and submitted via OpenReview. For novel work submissions, please add "[Novel]" to the paper title. For previously published work, add "[Published]" to the title.

A best paper award sponsored by McKinsey, KM-A and K4DP will be presented.

Important Dates

• Papers Due: Thursday, 27 July 2023
• Notification: Thursday, 31 August 2023
• Camera Ready: Thursday, 07 September 2023
• Workshop Date: 07 November 2023

1. RAG Architectures Leveraging KGs, LOD and Semantic Web Standards
2. Training and Fine-Tuning LLMs with Structured Data
3. Prompting Language Models with Structured Data
4. Evaluating RAG Architectures with Structured Data
5. Language Model and Ontology-Supported SPARQL Query Generation
6. Neurosymbolic Approaches for Integrating Language Models with LOD
7. Use Cases, Work-In-Progress and Bold Proposals for such RAG Systems