GitHub

Project Description: OpenSPG

What is the project about?

OpenSPG is a knowledge graph engine built upon the SPG (Semantic-enhanced Programmable Graph) framework. It's designed to represent, build, and reason over domain-specific knowledge graphs.

What problem does it solve?

It addresses the challenge of creating and using knowledge graphs in industrial settings, particularly where traditional RDF/OWL semantic complexity is impractical. It bridges the gap between big data and AI, enabling the transformation of raw data into actionable knowledge. It also solves the problem of knowledge graph evolution and iteration.

What are the features of the project?

  • SPG-Schema: Semantic modeling for property graphs, including subject, evolutionary, and predicate models.
  • SPG-Builder: Knowledge construction from both structured and unstructured data, integrated with big data systems. Includes operators for entity linking, concept standardization, and entity normalization.
  • SPG-Reasoner: Logical rule reasoning using a domain-specific language (KGDSL). Supports rule inference, neural/symbolic learning, and integration with large language models (LLMs).
  • KNext: A programmable framework for building knowledge graph solutions, offering extensible and user-friendly components.
  • Cloudext: A cloud adaptation layer for integrating with various graph storage/computation engines and machine learning frameworks.

What are the technologies used in the project?

  • Property Graph Model
  • RDF concepts (for semantic enhancement)
  • Big Data technologies (for data integration)
  • Natural Language Processing (NLP)
  • Deep Learning (including LLMs and Graph Learning)
  • Knowledge Graph Domain Specific Language (KGDSL)
  • Customizable/extensible graph storage and graph computation engines.
  • Customizable/extensible machine learning frameworks.

What are the benefits of the project?

  • Industrial Applicability: Overcomes the limitations of traditional semantic web technologies for practical, large-scale use.
  • Semantic Richness: Combines the simplicity of property graphs with the expressiveness of semantic web concepts.
  • Programmability: Allows for the definition of knowledge and logic rules in a machine-understandable way.
  • Evolvability: Supports continuous iteration and evolution of knowledge graphs.
  • Integration: Bridges big data, knowledge graphs, and AI technologies.
  • Extensibility: The programmable framework and cloud adaptation layer allow for customization and integration with various systems.
  • Open Source: Apache 2.0 License.

What are the use cases of the project?

  • Enterprise Supply Chain Knowledge Graph
  • Risk Mining Knowledge Graph
  • Medical Knowledge Graph
  • Any domain requiring the representation, construction, and reasoning over complex knowledge.
  • Knowledge-augmented generation (KAG) for boosting LLMs in professional domains.
  • Enterprise data interconnection.
openspg screenshot