#152 - Dave McComb - Knowledge Graphs, Semantics, and More
Nov 20, 2023
auto_awesome
Dave McComb, pioneer in knowledge graphs and semantics, discusses topics such as querying knowledge graphs, AI boom, waves of innovation, transitioning to semantic fashion, importance of ontology building, transparency in publicly traded companies, and flight preferences.
Combining large language models (LLMs) with knowledge graphs can enhance capabilities in various fields by leveraging the strengths of each.
The AI boom in the tech industry follows a pattern of hype, disillusionment, and enlightenment, indicating progress and maturity.
Implementing knowledge graphs involves gradually migrating use cases from relational databases, resulting in flexibility and comprehensiveness for integrated structured and unstructured data.
Deep dives
The importance of knowledge graphs and LLMs
Knowledge graphs and large language models (LLMs) are not fads but rather complementary to each other. LLMs are clever but can hallucinate and pose security risks, while knowledge graphs provide a non-probabilistic definitive statement of the best truth known. By combining the two, LLMs can be used to query knowledge graphs, and knowledge graphs can fact-check LLMs. This combination offers promising capabilities in various fields.
The cyclic nature of industry trends
Throughout history, various industries have witnessed cycles of innovation, hype, disillusionment, and enlightenment. This pattern repeats itself with each new promising technology or concept. The same is true for the current AI boom in the tech industry. While hype can lead to excessive expectations and subsequent disappointment, there is an underlying trend of progress and maturity in the industry.
The gradual implementation of knowledge graphs
Implementing knowledge graphs is a gradual process that often starts with bringing data from existing relational databases into a central, more flexible and interconnected structure. This process involves gradually eroding legacy systems by gradually migrating use cases to the knowledge graph. By preserving the distinctions and keeping the complexity manageable, organizations can benefit from the flexibility and comprehensiveness of knowledge graphs for integrated structured and unstructured data.
The benefits of a data-centered architecture
Shifting to a data-centered architecture involves placing data at the core and building application functionalities around it. This approach allows for better transparency, as data becomes the permanent element while applications are ephemeral. Data catalogs play a vital role in such architectures, providing metadata about different information domains and enabling users to navigate and access relevant data. A data-centered architecture offers more flexibility, transparency, and efficiency compared to traditional application-focused approaches.
The potential of revolutionizing data-centric accounting
Data-centric accounting represents a significant shift in the traditional approach to financial reporting. By focusing on the data as the core element, organizations can achieve immediate and accurate updates to their financial records, providing real-time insights and transparency. However, the adoption of data-centric accounting may face resistance from those invested in the current system of earnings management. Despite the challenges, the potential for improved transparency and accuracy makes this a compelling area for further exploration and innovation.
Dave McComb (Semantic Arts) is a pioneer in the use of knowledge graphs and semantics in data management. He joins to chat about these topics, and much more.
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
