NuKG Knowledge Platform
A nuclear knowledge graph platform accelerating R&D across the nuclear science community.
The Challenge
Navigating 94 Years of Neutron Science Literature
Nuclear science literature spans nearly a century — from Chadwick’s 1932 discovery of the neutron to today’s BNCT clinical trials. Over 100,000 papers cover neutron sources, cross-sections, isotope production, and fusion physics. Finding relevant prior work requires weeks of manual literature review. Existing tools like Google Scholar are general-purpose search engines. They cannot classify papers by neutron energy regime, extract structured nuclear data, or map citation networks within a specific domain.
Our Approach
AI-Powered Nuclear Knowledge Graph
NuKG is a multi-agent AI pipeline that autonomously builds a structured knowledge graph of neutron science. It crawls papers from OpenAlex and CrossRef, classifies relevance using Claude Haiku, extracts structured nuclear data using Claude Sonnet, synthesises relationships, and stores everything in PostgreSQL + Neo4j.
The production system has processed over 30,000 papers across three domain-specific pipelines, generating 62,000+ citation edges and structured extractions covering neutron flux values, cross-sections, facility types, and clinical trial phases. The platform runs at approximately $20/month.
Why It Matters
Key Benefits
Structured Nuclear Data
Every paper is analysed for neutron flux values, cross-sections, isotopes, facility types, and energy regimes — data that plain text search cannot extract.
Citation Network Analysis
62,000+ citation edges reveal research lineages, foundational papers, and emerging trends through citation velocity and clustering.
Domain-Specific AI Classification
Papers are classified by application area (BNCT, NAA, isotope production, NDE), neutron energy regime, and facility type — replacing weeks of manual review.
Who This Serves
Industries & Applications
- Nuclear R&D teams accelerating literature review and prior art discovery
- University research groups mapping citation networks in their specialisation
- Reactor and accelerator design teams finding published neutron spectra and cross-section data
- Regulatory bodies surveying the state of the art in nuclear application domains
Explore More
Related Applications
Get in Touch
Discuss NuKG Knowledge Platform
Interested in this capability? Tell us about your requirements and we'll explore how we can help.
Reach Us
Whether you're exploring collaboration, procurement, or research partnerships — we'd love to hear from you.