CoralX.AI: Multi-Hop, Redundancy-Aware Scientific QA via Hybrid Semantic-Graph Retrieval and RAG
Abstract
Coral reefs are some of the most important ecosystems on Earth, supporting a huge range of marine life and playing a crucial role in coastal economies. But they’re under serious threat from climate change and human activity. One of the biggest challenges in coral research today is the lack of clean, reliable data and the time-consuming process of finding and piecing together useful information. To help solve this, we built CoralX.AI, a custom-built, intelligent tool developed for the non-profit CoralX Foundation. It’s a specialized retrieval augmented generation (RAG) system that combines semantic search (using FAISS) with a graph-based knowledge network (built with NetworkX), designed to make coral-related research faster, smarter, and more accurate. The system breaks research documents into meaningful chunks using spaCy NLP, expands queries based on context, removes duplicates, and curates high-quality sources. It includes an APA-style citation generator to cite the sources. For answering questions, CoralX AI can perform exploratory deep, layered searches with subquery generation. In testing using the RAGAS evaluation framework, CoralX AI outperformed standard models, showing higher precision, relevance, and accuracy (up to 97% answer relevancy and 92% context precision). The system is currently live for users in its beta testing stage at coralx.ai.
Read the paper
Related studies
- Knowledge Management Perspective of Generative Artificial Intelligence
- Navigating Generative Artificial Intelligence Promises and Perils for Knowledge and Creative Work
- Characterizing the evolving research on enterprise content management
- Large language models and multimodal foundation models for precision oncology
- From Generative AI to Extended Reality: Multidisciplinary Perspectives on the Challenges, Opportunities, and Future of Educational Computing
- Generative Artificial Intelligence: A New Engine for Advancing Environmental Science and Engineering
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.