October 2024 Summaries

Neo4j has been recognized as a Strong Performer in The Forrester Wave: Vector Databases, Q3 2024, for its unique approach to improving GenAI application performance through GraphRAG, which combines RAG with knowledge graphs. This enables organizations to rapidly develop enterprise-grade GenAI applications, overcoming the limitations of vector-only RAG and accelerating GenAI adoption by providing enhanced context and accuracy, understanding complex data relationships, and better pattern recognition. GraphRAG-enhanced GenAI has numerous potential applications across industries, including recommendation engines, chatbots, fraud detection, and knowledge retrieval in healthcare and finance, making Neo4j a key player in the next generation of AI infrastructure.

The Neo4j GraphQL Library is undergoing significant changes with its roadmap for 2024-2025. Version 5, released in the first half of 2024, will transition to Long-Term Support (LTS) status until November 30, 2025, and continue to receive critical updates and security patches. This version will remain a stable choice for production use alongside Neo4j 4.x. The OGM (@neo4j/graphql-ogm) will also transition to LTS status but will no longer be in active development after October 2024. Version 6, released at the end of October 2024, will modernize the library and streamline its features, introducing changes such as a new subscription update engine and simplified configuration. This version acts as a transitional step for breaking changes between versions 5 and 7. The library's support timeline includes Version 5 until November 30, 2025, followed by active development and eventual release of Version 7 in early 2025. Users are advised to plan ahead and prepare for the upgrades, with recommendations to use Version 6 as a stepping stone between versions 5 and 7 or tolerate a breaking upgrade if possible.

Increased read throughput, better load distribution and higher availability can be achieved by distributing multiple Secondaries across availability zones in AuraDB. You can easily scale Secondaries as your workload requirements change, with options to manage them via the Aura Console or API. The Secondary count is retained when pausing and resuming a database, allowing for seamless setup maintenance without reconfiguration. However, it's essential to consider configuration limitations, including static Secondaries that don't support elastic scaling, and region restrictions, which require all instance and secondary components to be in the same cloud region.

Monitoring Neo4j performance is crucial for efficient data management. To efficiently monitor Neo4j and identify problematic queries, it's essential to understand the query optimization techniques and best practices. These include using EXPLAIN and PROFILE commands to analyze execution plans, leveraging domain knowledge of the data and network topology, and applying efficient traversal methods such as Quantified Path Patterns (QPP) or custom plugins. Subqueries can be used to break complex queries into smaller units, reducing memory allocation and improving efficiency. List comprehensions should be avoided due to potential performance issues. Aggregation operations should be minimized to avoid excessive memory usage. Combining multiple queries in a single query can reduce memory consumption and improve response time. Using Cypher queries instead of APOC plugins when possible can also optimize performance. Load testing with tools like JMeter can help identify and refine problematic queries, leading to improved availability and performance. By applying these techniques and best practices, it's possible to achieve 99.99 percent availability in production while demonstrating the resilience of Neo4j.

The GraphRAG Python package from Neo4j provides end-to-end workflows for creating knowledge graphs, retrieving data from them, and running retrieval-augmented generation (RAG) pipelines. It allows users to incorporate knowledge graphs into their applications, improving the relevance, accuracy, and explainability of RAG models. The package supports various retrievers, including vector retrievers that use Approximate Nearest Neighbor (ANN) search, Vector Cypher Retriever that combines vector search with graph traversal logic in Neo4j's Graph Query language, Hybrid Retriever that combines vector and full-text search, and Custom Retriever for tailored retrieval methods. The GraphRAG Python package is designed to be easy to use, even for those without extensive knowledge of Neo4j or RAG pipelines, providing comprehensive options for designing GenAI applications with knowledge graphs.

There's a new course available on GraphAcademy: Building Knowledge Graphs with LLMs, which teaches how to create and query knowledge graphs using large language models. This hands-on course uses the Neo4j LLM Graph Builder and Python to build knowledge graphs from unstructured data, allowing users to generate a source of truth for their unstructured data and ground GenAI applications using GraphRAG. The course is designed for advanced learners who have an understanding of Neo4j, integrating LLMs into applications, and Cypher, and upon completion, will have the skills to build a knowledge graph from their own unstructured data and use it to ground a GenAI chatbot.

Supply chain management is a crucial aspect of business operations that involves coordinating and overseeing every stage involved in the production and delivery of goods, from sourcing raw materials to delivering the final product to customers. Effective end-to-end supply chain planning ensures that the right products are available at the right time, in the right quantity, and at the right cost, optimizing both efficiency and customer satisfaction. Supply chains can be modernized on legacy systems by leveraging modern cloud-based systems and graph databases, which offer agility, scalability, and data-driven insights necessary for supply chain optimization. Graph technology optimizes supply chain management by visualizing complex supply chain networks, uncovering hidden relationships, enhancing decision-making with predictive analytics, and providing real-time data analysis for agility and responsiveness. Various industries, including agriculture, food supply chains, automotive, military logistics, and procurement optimization, are using graph technology to improve their supply chain operations, optimize processes, and make informed decisions that drive long-term success. The future of supply chain management and optimization will be shaped by advanced technologies like AI, blockchain, and IoT, along with a focus on sustainability, enabling businesses to predict disruptions, automate processes, and make smarter decisions.

AI is no longer just a breakthrough technology with tons of potential, but it's helping companies automate complex systems, improve decision-making, reduce expenses, and make processes safer, fairer, and better able to operate at scale. Business use of AI has become integral across industries, including healthcare, banking, fraud prevention, retail, and manufacturing, with spending on AI solutions expected to surpass $500 billion worldwide by 2027 and the market for AI technologies reaching $1.8 trillion by 2030. However, as AI evolves rapidly, it's difficult to keep up with the latest advances and ethical considerations, which is why creating and rolling out AI-aided apps involves considerable risk and companies are becoming aware of various negative ramifications if they don't handle the use of AI well. To mitigate these risks, companies need to create an AI code of ethics, be proactive in embracing principles, keep their data safe, design ethically, add context to their data for accuracy, transparency, and explainability, and adopt ethical AI practices with tools like Neo4j graph technology.

AI is a rapidly evolving technology that has become integral across various industries, including healthcare, banking, and manufacturing. However, as companies increasingly adopt AI to automate complex systems, improve decision-making, and reduce expenses, they must also be aware of the potential risks associated with its use, such as algorithmic bias, lack of accuracy, transparency, and explainability, and privacy violations. To mitigate these risks, it is essential for executives, data scientists, and developers to create an AI code of ethics, be proactive in implementing responsible AI practices, protect data privacy, design ethically, and add context to their data to ensure accuracy, transparency, and explainability. Companies can adopt ethical AI practices with Neo4j graph technology, which provides a framework for creating fair and trustworthy AI applications that reflect societal norms. By prioritizing AI ethics, companies can minimize risk, build trust, and promote robust and trustworthy outcomes.

Neo4j 5.26 LTS is coming with significant performance improvements in Cypher queries, including new operators, block format, parallel runtime, label inference, and eagerness analysis on logical plan. These features aim to enhance the query planner's ability to execute queries efficiently by leveraging advanced techniques such as pipelining, concurrent transactions, and eager operators. The parallel runtime supports read-only queries and can significantly speed up long-running analytical graph-global queries, while label inference continuously improves query planning by applying advanced deduction logic. Eagerness analysis on logical plan ensures that the query planner inserts eager operators where needed to optimize performance without sacrificing accuracy. Additionally, new Cypher language constructs such as quantified path patterns, shortest path, and vector indexes have been introduced to improve query performance in various ways. These features collectively contribute to faster and more efficient execution of Cypher queries in Neo4j 5.26 LTS.

This summary provides an overview of the text, which discusses the use of Large Language Models (LLMs) to create graphs from flat CSV files. The author, a developer experience engineer at Neo4j, attempts to use an LLM to model data in a graph and iteratively improves the process through fine-tuning and testing. The experiment involves loading a CSV file into a Pandas dataframe, analyzing each column, and generating Cypher statements to import nodes and relationships. The author also adds unique identifiers to the schema to ensure uniqueness of entities and creates constraints for the database. Finally, the author uses GraphCypherQAChain to query the graph and retrieve information about artists and tracks. The summary highlights the challenges faced during the experiment, including inconsistent JSON responses from the LLM and the need for fine-tuning the prompts. Overall, the text provides a practical example of using LLMs in graph data modeling and highlights potential applications and benefits of this approach.

Neo4j is a graph database that provides innovative solutions for identifying and mitigating fraud in real time through advanced entity link analysis. In an era where financial institutions face significant losses due to fraud, traditional methods are often not enough to combat increasingly sophisticated fraud schemes. Fraudsters are increasingly cunning, often collaborating in groups to exploit shared information and create fake identities, resulting in significant losses for banks and a challenging environment for fraud prevention teams. Neo4j's graph database can be used to write powerful Cypher queries that identify fraud rings by analyzing shared contact information, financial risks, and total credit limits associated with these rings. By leveraging the relationships inherent in data, financial institutions can significantly enhance their fraud detection capabilities, reduce losses, and improve their overall security posture.

Supply chain software is crucial for modern logistics, helping businesses tackle complexities and inefficiencies. With increasing investments in technology and a growing focus on transparency and ethical sourcing, supply chain management software plays a critical role in optimizing planning, sourcing, logistics, and operational efficiency. However, it's just one piece of the puzzle, as modern supply chains face constant pressure from disruptions, evolving customer demands, and increased requirements for visibility and ethics. To build more resilient systems, leaders are turning to complementary technologies like graph databases, which model complex relationships within supply chains, reveal hidden patterns, and enable real-time decision-making. Graph technology enhances supply chain resilience and agility by providing a dynamic approach that helps companies respond effectively to disruptions and manage intricate web of their supply networks.

GraphRAG is a card game that uses graph theory to improve generative AI chatbots, specifically Retrieval-Augmented Generation (RAG). The game simulates how a chatbot would answer questions about a topic by first spinning a spinner, which determines the level of information to provide. If it lands on STANDARD, the player answers the question directly; if RAG, they fetch relevant documents and use them to inform their answer; and if GRAPH RAG, they select a relationship type and draw all connected clue cards in one step. The game board represents a graph database, with clue cards as vertices and relationships between them as edges. GraphRAG improves chatbot performance by reducing hallucinations, improving explainability, and enabling answers about unseen topics.