Amazon Neptune ML now supports Real-time inductive inference
Starting with Amazon Neptune version 1.2.0.2, customers can now use real-time inductive inference with Amazon Neptune ML to enable machine learning (ML) predictions on nodes, edges and properties (entities) that were added to the graph after the ML model training process. With this launch, customers can make predictions on new data without requiring an update to their ML models. Amazon Neptune ML is powered by Amazon SageMaker, and uses Graph Neural Networks (GNNs), a machine learning technique purpose-built for graph that can improve the accuracy of most predictions for graphs by over 50% when compared to making predictions using non-graph methods based on published research from Stanford University.
Starting with Amazon Neptune version 1.2.0.2, customers can now use real-time inductive inference with Amazon Neptune ML to enable machine learning (ML) predictions on nodes, edges and properties (entities) that were added to the graph after the ML model training process. With this launch, customers can make predictions on new data without requiring an update to their ML models. Amazon Neptune ML is powered by Amazon SageMaker, and uses Graph Neural Networks (GNNs), a machine learning technique purpose-built for graph that can improve the accuracy of most predictions for graphs by over 50% when compared to making predictions using non-graph methods based on published research from Stanford University.
Customers often need real-time predictions for use cases like fraud detection, product recommendations, and identity resolution. For example, when a new user attempts to create an account on an e-commerce platform, the business may want to generate a fraud prediction score using machine learning and take risk mitigation actions like blocking the account creation or holding it for manual review. With real time inductive inference for Neptune ML, customers can get near real-time predictions on new data by using existing Neptune ML models without retraining their ML models each time. Additionally, customers can now train and deploy Neptune ML models faster and save costs by training on a representative sample of their graph data and then deploying it to make predictions on any entity in the graph.
To learn more about real-time inductive inference, see the Neptune ML documentation page. To get started with Neptune ML, use this CloudFormation quick-start and walk through pre-built Neptune ML notebook tutorials that are included in the Neptune Notebook.
For more information on pricing and region availability, refer to the Neptune pricing page and AWS Region Table. There are no additional charges for using Amazon Neptune ML real-time inductive inference. You only pay for the resources provisioned such as Amazon Neptune, Amazon SageMaker, Amazon CloudWatch, and Amazon S3.