HighlightGoogle Cloud

Knowledge Catalog: Gemini context graph for agents and governance

Google Cloud overview of Knowledge Catalog (formerly Dataplex Universal Catalog): an AI-powered catalog that builds a context graph to ground agents.

As of April 10, 2026, Dataplex Universal Catalog is renamed Knowledge Catalog while API, client library, CLI, and IAM names stay the same. The product is described as a Gemini-powered catalog that extracts semantics from structured and unstructured data into a dynamic context graph for discovery, policy, and agent grounding to reduce hallucinations.

Based on: Knowledge Catalog overview | Google Cloud Documentation · Google Cloud

HighlightNeo4j

Neo4j graph modeling ties domain questions to storage shape

Neo4j getting-started overview of graph data modeling steps from domain use cases through test, Cypher load, and refactor.

Graph data modeling defines query logic and stored structure; a well-designed model improves performance, flexibility, and storage use. The process covers understanding the domain and use-case questions, extracting entities and relationships, testing against an initial model, loading test data with Cypher, measuring performance, and refactoring as use cases change.

Based on: What is graph data modeling? - Getting Started · Neo4j

HighlightNeo4j

Neo4j MERGE: match-or-create patterns with ON MATCH/ON CREATE

Neo4j Cypher manual for MERGE, which binds existing patterns or creates missing ones, with constraint guidance.

MERGE combines MATCH and CREATE: if the exact pattern exists it binds like MATCH; otherwise it creates like CREATE. ON MATCH and ON CREATE allow different actions for each case. The manual recommends creating constraints before merging for index-backed performance and to prevent unintended divergent data.

Based on: MERGE - Cypher Manual · Neo4j

HighlightW3C Shape Expressions Community Group / shex.io

ShEx 2.1 defines shapes and node constraints for describing RDF graph structure

Final Community Group Report (8 Oct 2019) for Shape Expressions Language 2.1: RDF node and graph structure descriptions for validation and interfaces.

Shape Expressions (ShEx) describe RDF nodes and graph structures: node constraints on IRIs, blank nodes, or literals, and shapes over triples with predicates, cardinalities, and datatypes. Shapes can communicate structures for processes or interfaces, generate or validate data, or drive user interfaces. ShEx 2.1 adds IMPORTS and language tag value sets.

Based on: Shape Expressions Language 2.1 · W3C Shape Expressions Community Group / shex.io

HighlightStanford Center for Biomedical Informatics Research

Protégé is Stanford’s free OWL 2 editor for desktop and collaborative web work

Protégé homepage: open-source OWL ontology editor (Desktop and WebProtégé), used on OBO Foundry, WHO ICD-11, and NCI Thesaurus.

Protégé is a free, open-source OWL ontology editor from Stanford, available as Protégé Desktop and WebProtégé. It supports the OWL 2 lifecycle from modelling through reasoning, querying, and collaboration. Notable uses include OBO Foundry ontologies, WHO ICD-11 development, and the NCI Thesaurus.

Based on: Protégé · Stanford Center for Biomedical Informatics Research

HighlightApache Software Foundation

Apache Jena’s jena-shacl runs W3C SHACL Core and SPARQL constraints in practice

Jena documentation for jena-shacl: SHACL Core and SPARQL constraints, compact syntax, CLI validate/parse, and Fuseki endpoint integration.

jena-shacl implements W3C SHACL Core and SHACL SPARQL Constraints, plus compact syntax and SPARQL-based targets. The shacl CLI validates and parses shapes; Fuseki can expose a fuseki:shacl operation that accepts a posted shapes graph and optional graph/target parameters.

Based on: Apache Jena - Apache Jena SHACL · Apache Software Foundation

HighlightMorgan & Claypool (open online edition)

A 2018 handbook on RDF validation covering ShEx, SHACL, and data quality

Open HTML edition of Validating RDF Data (Morgan & Claypool, 2018) by Labra Gayo, Prud’hommeaux, Boneva, and Kontokostas.

Validating RDF Data is a Synthesis Lectures book on Semantic Web theory and technology. The contents move from RDF and data quality through Shape Expressions and SHACL, including shapes, constraints, and related tooling. A free HTML edition is maintained with errata and example source.

Based on: Validating RDF Data · Morgan & Claypool (open online edition)

HighlightJSON-LD CG / Digital Bazaar community site

JSON-LD turns ordinary JSON into Linked Data that can cross site boundaries

Project site for JSON-LD: a JSON-based Linked Data format, W3C specs, playground, and conforming libraries across many languages.

JSON-LD is a lightweight Linked Data format based on JSON, meant to be readable by humans and usable in programming environments, REST services, and document databases. Linked Data here means standards-based, machine-readable data that can follow links across sites. The site points to W3C recommendations, a playground, and conforming implementations.

Based on: JSON-LD - JSON for Linked Data · JSON-LD CG / Digital Bazaar community site

Highlightdbt Labs

dbt entities are typed join keys that link semantic models

dbt docs define entities as business concepts used as join keys—primary, unique, foreign, or natural—in the Semantic Layer.

Entities represent real-world concepts such as customers or transactions and act as join keys across semantic models in dbt’s Semantic Layer (v1.12+). Required parameters are name and type; types are primary, unique, foreign, and natural. Names must be unique within a model and may use expr; entities can also be used as dimensions.

Based on: Entities | dbt Developer Hub · dbt Labs

Highlightdbt Labs

dbt semantic models are MetricFlow graph nodes tied to DAG models

dbt docs explain semantic models as YAML-configured nodes and entities that MetricFlow uses for the Semantic Layer.

In dbt v1.12+, semantic models are the foundation for MetricFlow’s Semantic Layer: nodes linked by entities in a semantic graph, annotated on dbt models for metric use. Each DAG model maps to one semantic_model YAML block (or Apache Ossie docs); components include name, time dimension, and entities.

Based on: Semantic models | dbt Developer Hub · dbt Labs

HighlightarXiv

Increasing the LLM Accuracy for Question Answering: Ontologies to the Rescue!

Paper on improving question answering systems with large language models using ontologies.

This paper presents an approach to improve the accuracy of question answering systems with large language models by leveraging ontologies. The authors propose a method that consists of ontology-based query check and LLM repair, which increases the overall accuracy to 72%. The results provide further evidence that investing knowledge graphs, namely the ontology, provides higher accuracy for LLM-powered question-answering systems.

Based on: Increasing the LLM Accuracy for Question Answering: Ontologies to the Rescue! · arxiv.org

Highlightgit.nextgraph.org

oxigraph

A Rust-based graph database implementing the SPARQL standard.

Oxigraph is a graph database written in Rust that implements the SPARQL standard. It provides a compliant, safe, and fast graph database based on the RocksDB key-value store. Oxigraph also includes utility functions for reading, writing, and processing RDF files.

Based on: oxigraph · git.nextgraph.org