Sydney, Australia – Progress, a software development firm, has announced that it has recently launched the early-access release of Progress MarkLogic Server 12. This latest release builds on the platform’s robust features by adding native vector support and new relevance-based algorithms, providing advanced capabilities for customers to leverage AI across enterprises.
The native vector support enables large-scale similarity searches for AI systems on vast amounts of unstructured data. It also natively performs vector operations in MarkLogic Server to improve GenAI retrieval accuracy.
Another notable addition is the BM25 relevance ranking, which improves recall with better document ranking to return the most relevant search results at the top. This feature allows combining best-match and vector search to rerank results for optimal search relevance.
With MarkLogic Server comprehensive search, organisations can now combine queries, merge result scores, and re-rank them with a single expression via its unified Optic API that parses the content most relevant to a user query to the LLM model.
Rounding out this suite of new features is the ability to graph the shortest path, significantly solving real-world problems related to route planning, network optimisation, and resource allocation. This capability enhances contextual relevance of results and response time in AI-enhanced recommendation systems through semantic searches against knowledge graphs or semantic data in graph RAG architectures.
Through the enhancements, customers can now seamlessly integrate generative AI into their enterprise applications using a flexible retrieval-augmented generation framework, enabling securely ground LLM models in their private data to utilise fact-based query responses generated from AI models.
These new updates further enhance the relevance and accuracy of GenAI responses through improved search techniques.
For instance, a single API allows users to define, orchestrate, and re-rank search scores for faster, more flexible, and more secure graph RAG implementation. This RAG architecture matches the data model and metadata maturity, allowing greater use of enterprise data for GenAI-enhanced applications.
Complementing these enhancements, the company has also unveiled its latest MarkLogic® Flux, an application designed to add improved data movement and transformation capabilities to the core functionality of the MarkLogic platform.
This new platform, which includes MarkLogic Server, merges multi-model data management with real-time, relevance-based search and semantic capabilities to provide an adaptable, secure foundation for an organisation’s GenAI solutions. This move leverages GenAI in their enterprise with a flexible RAG architecture that augments LLMs with domain-specific knowledge to increase response accuracy, democratise access to information across the organisation, and power AI-enhanced applications.
John Ainsworth, EVP and general manager, application and data platform at Progress, said, “As AI becomes increasingly integrated into the enterprise, it is crucial to not only improve the precision of the results delivered but also to safeguard the enterprise’s valuable data.”
He further noted, “With the latest release of MarkLogic Server, our customers have access to a dependable database management system capable of generating top-quality AI responses while also protecting their reputation and intellectual property investment.”
