Latest Neo4j Graph Platform Reveals Context for AI Applications using a Connections-First Approach
Neo4j, the market leader in connected data, announced today the upcoming release of Neo4j 3.5, the native graph platform designed to drive the success and adoption of real-time business applications, including artificial intelligence (AI) and machine learning (ML) systems.
Neo4j customers – including eBay and Caterpillar – have demonstrated that connected graph datasets are a foundational element of enterprise AI applications. Graph-based data models provide the necessary context for AI applications by capturing facts related to and relationships among people, processes, applications, data and machines.
Informed by successful AI customer deployments – including knowledge graphs, fraud detection, recommendation systems and conversation engines – Neo4j 3.5 delivers foundational features for AI-powered systems of connection to generate bottom-line business value.
Adrian Bowles, founder of STORM Insights, had this to say about the synergy between artificial intelligence and graphs:
“The way we organize and represent knowledge in AI-powered systems has a profound effect on what and how they can learn,” said Bowles. “The essence of cognition is the ability to analyze or synthesize what we know, within context, using relationships between existing units of knowledge to enable the synthesis of new knowledge. Representing these relationships in a graph enables more efficient and effective artificial cognitive processes.”
Data Relationships Drive Context for AI
Most current models and techniques that underpin AI systems are not optimized for detecting connections or traversing relationships within datasets. Property graphs reveal and navigate connections, and therefore discover context by linking attributes and complex relationships across the graph, making them the ideal data structure to power machine learning models.
Furthermore, the robustness of the Neo4j Graph Platform and its breadth of use cases demonstrate that it functions as the ideal system of record to safeguard and maintain connected data even while complex algorithms are being run multiple times per second.
A perfect example of Neo4j being deployed within a machine learning context is the German Center for Diabetes Research (DZD).
“Neo4j enables a new dimension of data analyses to fight diabetes by helping us to connect highly heterogeneous data from various disciplines, species and locations to build an invaluable body of knowledge,” said Dr. Alexander Jarasch, Head of Bioinformatics and Data Management at the DZD. “By applying modern machine learning techniques to our Neo4j graph, we are getting closer to understanding this complex disease to help diabetics and those with prediabetes.”
Neo4j CEO and Co-Founder Emil Eifrem notes that the forthcoming Neo4j 3.5 release will power the evolution of AI systems even more effectively in the future.
“Our customers are pushing the envelope of what can be achieved with graph-powered AI, which we think of as intelligent systems of connection,” said Eifrem. “They are able to do so because graph technology fundamentally embraces relationships as first-class entities. This relationships-first approach adds context to data, which is key to accurate, well-informed predictions. With Neo4j 3.5 we have worked extensively with our customers to deliver the robustness and scale they need for tomorrow’s graph-powered artificial intelligence systems.”