⭐ Help us reach more developers and grow the StrategyKG community. Star this repo!
StrategyKG is an open-source knowledge graph designed for strategic AI applications. It provides meaningful knowledge representations and tools to enhance strategic planning and decision-making, enabling AI systems to offer more informed and effective strategic insights.
Knowledge graphs are structured representations of knowledge that capture the relationships between entities. They serve as powerful tools for organizing and querying complex information, making it easier to understand and utilize.
Developing strategic-level applications is highly difficult, requiring AI systems to provide specialized advice and solve intricate problems. Large language models, while powerful, need to be augmented with knowledge graphs to enhance their capabilities in delivering domain-specific recommendations and addressing complex issues. StrategyKG is designed to support the development of strategic AI applications by providing a rich and structured knowledge base.
Strategy is complex reasoning, and both symbolic representation and vector representation find it difficult to directly describe the application relationships of strategies. Based on the "abstraction ladder" in cognitive science, StrategyKG defines four levels of strategic knowledge, making strategic knowledge modeling three-dimensional to further explore the application of complex reasoning. Each level will provide specific content under different licensing agreements to balance open-source and commercial interests, allowing them to complement each other.
| Level | Abstraction | Knowledge Type | Temporal | License | |
|---|---|---|---|---|---|
| Level 0 | Conceptual Layer | High-level Abstraction | Theoretical Knowledge | Atemporal | CC0 (Public) |
| Level 1 | Categorical Layer | Mid-level Abstraction | Domain Knowledge | Long-term Trends | CC0 (Public) |
| Level 2 | Entity Layer | Low-level Abstraction | Factual Knowledge | Historical Tense | CC BY 4.0 (Attribution Public) |
| Level 3 | Dynamic Layer | Concrete Instances | Procedural Knowledge | Real-time Stream | Commercial Use (Requires License) |
Building a knowledge graph is a complex and long-term task that requires a large number of accurate documents and effective datasets. To meet the needs of different users and promote the growth of StrategyKG, we use the capability maturity model to measure the development path of strategic AI applications based on knowledge graphs (and large language models). The Strategic AI Application Capability Maturity Model includes four levels, each targeting different user needs and providing specific content under different licensing agreements.
| Maturity Level | 0 - Starter | 1 - Learner | 2 - Executor | 3+ - Strategist |
|---|---|---|---|---|
| Suitable Users | Newcomers to strategic planning | Individuals seeking to expand their strategic knowledge | Practitioners implementing and monitoring strategies | Experienced strategists requiring complex strategic planning and innovation |
| Content Provided | Basic concepts, simple frameworks, and introductory case studies | Advanced frameworks, industry-specific analyses, and strategy formulation tools | Strategic optimization methods, competitive analysis, and risk management tools | Strategic innovation methods, complex models, global strategy analyses, and decision support systems |
| License Agreement | CC0 (Public Domain) | CC0 (Public Domain) | CC BY 4.0 (Attribution Required) | Commercial Use (Contact for Licensing) |
To get started with StrategyKG, simply clone the repository and explore the resources available at each maturity level. We welcome contributions from the community to enrich the knowledge graph and support strategic AI applications.
We encourage contributions to StrategyKG. Whether you're a newcomer or an experienced strategist, your insights and expertise can help enhance this knowledge graph. Please refer to our Contribution Guidelines for more information.
If you are interested in a business partnership with this open-source project, please contact [email protected].