Internship: Maintenance and Real-Time Updating of Deployed Knowledge Graphs
Description
(This project is also available as an MSc Thesis)
Company: Marel
Location: Boxmeer
Background
Knowledge Graphs have emerged as a powerful tool for representing vast amounts of interconnected data. By structuring data in a graph format, enterprises can uncover relationships and insights that are often hidden in traditional databases. This approach not only enhances data accessibility and integration but also supports advanced analytics and decision-making processes.
However, maintaining and updating Knowledge Graphs presents significant challenges. As enterprise data continuously evolves, ensuring the accuracy and relevance of the Knowledge Graph requires robust mechanisms for data ingestion, validation, and synchronization. Inconsistent data sources, varying data formats, and the sheer volume of updates can lead to data integrity issues and increased complexity in graph management. Addressing these challenges is crucial for leveraging the full potential of Knowledge Graphs in enterprise environments.
Large companies often manage vast amounts of both structured data, such as product lifecycle management data, and unstructured data, including user manuals, service manuals, and technical reports. Knowledge Graphs (KG) offer a solution by integrating these diverse data sources into a single, cohesive database, facilitating easier querying and analysis.
Problem description
In this project, we are constructing a knowledge graph from company data. The graph is not connected in real-time to the data sources, necessitating continuous updates with new or changing information. It is crucial to develop a method for maintaining and updating the knowledge graph to ensure users have access to the most current information.
Therefore, this project works on developing a system for the real-time maintaining and updating of deployed knowledge graphs. By implementing automated processes to continuously sync the knowledge graph with its data sources, the system will ensure that the graph remains up-to-date. This in-turn should enhance KG reliability and usefulness, providing users with timely and accurate information for their queries and analyses.
Goals
- Investigate the needs and characteristics for knowledge graph updating in an industrial setting
- Demonstrate knowledge graph updates on a local level (only concerning enties, not ontology) at either continuous or periodic intervals
- Assess whether the speed at which this is possible is sufficient for end-users
- Explore methods for verifying new data and consistency with pre-existing KG data
Student requirements
- Having passed (or intending to take) Knowledge Engineering (2AMD20) is a pre
- Experience with graph databases and graph query languages (such as Cypher) is a pre
- Pragmatic but research oriented, with the intention to publish the results
- Work in Boxmeer 3-4 days a week
Perks
- An internship allowance and travel allowance
Details
- Supervisor
-
Nick Yakovets
- Company
- Marel
- Interested?
- Get in contact
- Link
- Thesis