D.2.14. Large language model AI¶
Leveraging the power of LLM AI
The capability Large language model AI (D.2.14) is part of the capability area Technology Execution in the Technology Pillar.
Leveraging the power of LLM AI
While an Enterprise Knowledge Graph (EKG) can be a valuable resource for fact-checking, it is important to note that the EKG is primarily designed to organize and connect structured and unstructured data sources within an organization. It serves as a knowledge repository and enables efficient data discovery and access.
In the context of fact-checking a large language model like ChatGPT, the EKG can contribute in the following ways:
- Data Sources and References: The EKG can integrate various data sources---and Data Products---that contain factual information, including authoritative databases, trusted repositories, published research, and validated sources. These sources can be linked and associated with specific facts or claims, providing a reference point for fact-checking.
- Semantic Relationships: The EKG captures semantic relationships between entities, concepts, and facts. By leveraging these relationships, fact-checking can be performed by cross-referencing claims made by the language model with the connected information within the EKG. Semantic relationships help establish context, relevance, and coherence, enabling a more comprehensive evaluation of the facts.
- Collaborative Validation: The EKG can facilitate collaborative validation and verification processes by involving subject-matter experts, domain specialists, and other stakeholders. These experts can contribute their knowledge, insights, and expertise to validate the claims made by the language model, identify potential inaccuracies, and provide additional context to ensure the accuracy and reliability of the information.
- Knowledge Representation: The EKG's structured knowledge representation allows for the organization and categorization of factual information. By leveraging the EKG's taxonomy, ontology, and metadata capabilities, fact-checking efforts can be streamlined, enabling efficient search, retrieval, and analysis of relevant information for validation.
- Metadata and Provenance: The EKG can store metadata and provenance information about the sources, authors, and credibility of data. This metadata can serve as a basis for evaluating the trustworthiness and reliability of the information used by the language model. It helps identify potential biases, assess the quality of sources, and ensure transparency in the fact-checking process.
It's important to note that while an EKG can provide valuable resources for fact-checking, it should be complemented with other fact-checking methodologies and tools. Human expertise, independent verification, and critical thinking are essential components in ensuring accurate and reliable information.
Combining the power of an EKG with human judgment and expert input can enhance the fact-checking process, providing a comprehensive and reliable framework for evaluating the claims made by large language models like ChatGPT.
Warn
Work in progress, describe the various measurable dimensions and concerns related to this capability
Warn
Work in progress. Describe the five levels of maturity for this Capability.
Warn
Work in progress. Explain how EKG contributes value and how this capability or capability- enables higher levels of maturity for the EKG (which in turn provides more value to the business)
Warn
Work in progress, describe how this capability is possibly being delivered today in a non-EKG context and optionally what the issues are that EKG could or should improve
Warn
Work in progress, describe how this capability would be delivered or supported using an EKG approach, making the link to the "how" i.e. the EKG/Method.
Warn
Work in progress, list examples of use cases that contribute to this capability, making the link to use cases in the catalog at https://catalog.ekgf.org/use-case/..