LLM+KG

LLM+KG:

Large Language Models (LLMs), e.g., ChatGPT and LLaMA are revolutionizing the fields of artificial intelligence and natural language processing (NLP). Recent LLMs browse Web knowledge and learn from external sources, warranting the coupling of knowledge graphs (KGs) and LLMs. The possibility of bridging KGs with LLMs has attracted increasing interest in the area of knowledge engineering. On one hand, LLMs can be enhanced with KGs to provide answers with more contextualized facts. On the other hand, downstream tasks, e.g., KG curation, embedding, and search can also benefit by adopting LLMs. It remains an interesting direction to explore effective interactions between LLMs and KGs, where many recent advances come from deep learning, information retrieval, NLP, and computer vision domains. The workshop, titled “LLM+KG: Data Management Opportunities in Unifying Large Language Models + Knowledge Graphs”, is targeted for data management researchers, aiming to discuss interesting opportunities such as data cleaning, modeling, designing of algorithms and systems, scalability, fairness, privacy, usability, explainability, and etc.

We solicit unpublished papers discussing issues and successes under the broad category of LLM-enhanced KGs, KG-enhanced LLMs, and unifying LLMs + KGs in the following areas (and beyond):

• KG-enhanced Pre-training of LLMs

• KG-enhanced Fine-tuning of LLMs

• KG-enhanced Inference of LLMs

• KG-enhanced Validation and Explainability of LLMs

• LLM-enhanced KG Creation

• LLM-enhanced KG Completion

• LLM-enhanced KG Embedding

• LLM-enhanced KG Querying

• LLM-enhanced KG Analytics

• LLM-enhanced Domain-specific KG Applications

Additionally, the paper must have a clear data management focus, e.g., discussing data management solution(s) such as (but not limited to):

• Data and Input Modeling for LLMs+KGs

• Data Cleaning, Integration, and Augmentation with LLMs+KGs

• Multi-modal Data Management with LLMs+KGs

• Vector Data Management for LLMs+KGs

• Accuracy and Consistency of LLMs+KGs

• Efficiency and Scalability of LLMs+KGs

• Bias and Fairness with LLMs+KGs

• Explainability and Provenance of LLMs+KGs

• Usability of LLMs+KGs

• Security and Privacy for LLMs+KGs

• Optimizing KG Databases and Systems with LLMs

• Empirical Benchmark and Ground Truth in Emerging Applications with LLMs+KGs

Submission Details:

We solicit and select three types of papers

• Survey Papers/Tutorials: these papers survey the related work in specific sub-areas and lay out the agenda for future work.

• New/ Late-breaking Results: these papers report the newest preliminary results about the most promising problems in the field.

• Vision Papers: these papers are devoted to discussing problems that we face currently and anticipate for the future.

We welcome the papers that fall under short papers of at most 9 pages and long papers up to 18 pages, including bibliography. Submissions must adhere to CEUR-WS formatting guidelines with 1-column style available at: http://ceur-ws.org/Vol-XXX/CEURART.zip. An Overleaf page for LaTeX users is available as template at: https://www.overleaf.com/read/xztwvxtwbzrn#ac9ca2. All submissions must be submitted in PDF through: https://cmt3.research.microsoft.com/LLMKG2024/. Submissions will be reviewed in a single-blind manner, and all author names and affiliations should be included. Papers that do not follow the guidelines or are not within the scope of relevant topics will be desk rejected. We also expect that publications from DB venues, e.g., SIGMOD/VLDB/ICDE/EDBT etc. are cited.

Submissions will be reviewed by at least three members of the Program Committee. All accepted papers will be published online via CEUR-WS. The workshop will be in-person and at least one author of each accepted paper is required to register. Best papers will be invited to submit extensional versions to the special issue: Neuro-Symbolic Intelligence: Large Language Model Enabled Knowledge Engineering in the World Wide Web Journal, and the Data Intelligence Journal.