Amazon Neptune ML

Easy, fast, and accurate predictions for graphs

Overview

Amazon Neptune ML is a new capability of Neptune that uses graph neural networks (GNNs), a machine learning (ML) technique purpose-built for graphs, to make easy, fast, and more accurate predictions using graph data. With Neptune ML, you can improve the accuracy of most predictions for graphs by over 50% when compared to making predictions using nongraph methods.

Making accurate predictions on graphs with billions of relationships can be difficult and time-consuming. Existing ML approaches such as XGBoost can’t operate effectively on graphs because they are designed for tabular data. As a result, using these methods on graphs can take time, require specialized skills from developers, and produce suboptimal predictions.

The Deep Graph Library (DGL), an open source library to which AWS contributes, makes it easier to apply deep learning to graph data. Neptune ML automates the heavy lifting of selecting and training the best ML model for graph data and lets users run ML on their graph directly using Neptune APIs and queries. As a result, you can now create, train, and apply ML on Neptune data in hours instead of weeks without the need to learn new tools and ML technologies.

ML and generative AI

Neptune ML automatically creates, trains, and applies ML models on your graph data. It uses DGL to automatically choose and train the best ML model for your workload so that you can make ML-based predictions on graph data in hours instead of weeks.

Neptune ML uses GNNs, a state-of-the-art ML technique applied to graph data that can reason over billions of relationships in graphs, so that you can make more accurate predictions.

*Neptune ML uses GNNs to make predictions that can be more than 50% more accurate than nongraph ML, based on published research from Stanford University.

LangChain is an open source Python framework designed to simplify the creation of applications using large language models (LLMs). Neptune integration with LangChain allows developers to use LangChain’s open source framework to simplify the creation of context-aware applications.

With Neptune and LangChain, you can return a response based on the provided context and query a Neptune graph database using the openCypher query language. For example, you can use the Neptune openCypher QA Chain to translate English questions into openCypher queries and return a human-readable response. This chain can be used to answer questions such as “How many outgoing routes does the Austin airport have?”

For more details about the Neptune openCypher QA Chain, visit the open source LangChain documentation.

LlamaIndex is a open-source data framework for connecting custom data sources to large language models (LLM) and supports using knowledge graphs with LLMs.

With LlamaIndex, can use Neptune as a graph store or a vector store to build generative AI applications using techniques like GraphRAG.

Use cases

Companies lose millions (even billions) of dollars in fraud and want to detect fraudulent users, accounts, devices, IP addresses, or credit cards to minimize the loss. You can use a graph-based representation to capture the interaction of the entities (user, device, or card) and detect aggregations such as when a user initiates multiple mini transactions or uses different accounts that are potentially fraudulent.

Learn more

An identity graph provides a single unified view of customers and prospects based on their interactions with a product or website across a set of devices and identifiers. Organizations use identity graphs for real-time personalization and advertising targeting for millions of users. Neptune ML automatically recommends next steps or product discounts to certain customers based on characteristics such as past search history across devices or where they are in the acquisition funnel.

Learn more

Knowledge graphs consolidate and integrate an organization’s information assets and make them more readily available to all members of the organization. Neptune ML can infer missing links across data sources and identify similar entities to enable better knowledge discovery for all.

Learn more

Traditional recommendations manually use analytics services to make product recommendations. Neptune ML can directly identify new relationships on graph data and easily recommend the list of games a player would be interested to buy, other players to follow, or products to purchase.

Pricing

There are no upfront investments needed. You only pay for the AWS resources used such as Amazon SageMaker, Neptune, and Amazon Simple Storage Service (Amazon S3).

Getting started

The easiest way to get started with Neptune ML is to use the prebuilt AWS CloudFormation quick-start templates. You can also walk through the Neptune ML notebooks to see end-to-end examples of node classification, node regression, and link prediction using the prebuilt CloudFormation stack.

Create a Neptune ML stack