Experts at Enterprise Knowledge break down knowledge graphs and machine learning
From automated fraud detection and intelligent chatbots, to dynamic risk analysis and content-based recommendation engines, knowledge graphs coupled with machine learning are becoming the go-to solution as enterprises hunt for more effective ways to connect the dots between the data world and the business world.
KMWorld held a webinar featuring Bess Schrader, senior consultant, Enterprise Knowledge and Joe Hilger, CTO, Enterprise Knowledge, who discussed how to get started with knowledge graphs and machine learning to achieve Enterprise AI.
A knowledge graph is a specialized graph of the things we want to describe and how they are related, Schrader and Hilger explained. Graphs are excellent for modeling language and meaning.
The structure and features of the network itself serve as the foundation for a variety of ML processes including classification, regression, clustering, and link or node prediction, they said.
Machine Learning is one type of artificial intelligence system, in which a machine “learns” from your data. It is typically an integrated system that can perform actions that traditionally require human intelligence.
AI is frequently a solution in search of a problem and many companies try to buy a product that solves the problem or tackle too much all at once. Schrader and Hilger offered a simple, iterative approach that allows organizations to see success in months not years.
The roadmap to Enterprise AI includes tips such as:
- Define vision and relevant use cases
- Inventory and organize priority knowledge, content, and data
- Map relationships
- Conduct a proof of concept
- Automate, optimize, and scale
Define overarching vision that outlines a clear meaning and business value of AI for your enterprise. Select use cases that support future implementations. Relevant use cases include natural language search, content and data governance and discovery, and compliance and operational risk prediction.
Taxonomy, metadata, and data catalogs allow for effective classification and categorization of both structured and unstructured information for findability. When inventorying millions of content items, use tools to automate the process. Lay the groundwork for getting your information into a machine-readable format.
Ontologies map the relationships and connections existing between information and data components (both structured and unstructured). Increase discoverability of hidden content and information to optimize search experience. Lay foundations for intelligent AI capabilities, like text mining and context-based recommendations.
Start small in a test environment to iteratively validate the data model against real data to quickly show progress. Enhance your knowledge graph by tagging internal and external sources of information.
Develop a prioritized backlog to incrementally prove and deliver on Enterprise AI initiatives, like semantic search, optimized data management, and governance. Iterate and scale with each new business question and data source.
An archived on-demand replay of this webinar is available here.