Advances in Causal Discovery and Knowledge Graph Construction

The field of artificial intelligence is moving towards more advanced and efficient methods for causal discovery and knowledge graph construction. Researchers are leveraging large language models (LLMs) and retrieval-augmented generation techniques to improve the accuracy and scalability of these methods. The use of LLMs is allowing for more efficient causal graph construction, while retrieval-augmented generation is enabling the incorporation of external knowledge into the process. This is leading to more accurate and interpretable results, and is being applied to a range of domains, including natural language processing and computer vision. Notable papers in this area include: Fairness-Driven LLM-based Causal Discovery with Active Learning and Dynamic Scoring, which introduces a framework for fairness-driven causal discovery using LLMs. CausalRAG: Integrating Causal Graphs into Retrieval-Augmented Generation, which proposes a novel framework that incorporates causal graphs into the retrieval process. SLIDE: Sliding Localized Information for Document Extraction, which introduces a chunking method that processes long documents by generating local context through overlapping windows.

Sources

Fairness-Driven LLM-based Causal Discovery with Active Learning and Dynamic Scoring

SLIDE: Sliding Localized Information for Document Extraction

Retrieval Augmented Generation and Understanding in Vision: A Survey and New Outlook

WikiAutoGen: Towards Multi-Modal Wikipedia-Style Article Generation

Quantifying Symptom Causality in Clinical Decision Making: An Exploration Using CausaLM

CausalRAG: Integrating Causal Graphs into Retrieval-Augmented Generation

D4R -- Exploring and Querying Relational Graphs Using Natural Language and Large Language Models -- the Case of Historical Documents

Leveraging Large Language Models for Automated Causal Loop Diagram Generation: Enhancing System Dynamics Modeling through Curated Prompting Techniques

Built with on top of