Saffron's software solutions leverage a highly flexible, machine-learning engine, and are developed, and deployed using a common architecture called SaffronEnterprise™.

The memory-based engine captures and stores associations using the same basic representations that form the core of human memory. It is based on a patented associative memory technology with a 100% Pure Java™ Application Programming Interface (API) that is easily embedded into existing applications and systems, empowering them with highly adaptive, intelligent functionality.

Unlike traditional neural networks and data mining, a memory-based engine supports agent-based learning. Agents that learn in the real world must be very quick to learn, sometimes with little data, incrementally case-by-case, and without parameter tweaking by a knowledge engineer. These agents learn by merely observing, as do humans. Importantly, the Saffron engine can handle many data types, including structured, semi-structured, and unstructured data.

Saffron's core machine learning engine provides the data scaling and system response time that truly differentiates Saffron's entire suite of products.

SaffronEnterprise™ development architecture includes a standard set of tools that enable Java developers to quickly and easily take their source data and integrate it into their application.

SaffronEnterprise™ allows applications to have fully supported multi-agent systems, which are required for enterprise scale deployment. It is based on an entity network system where each entity has its own memory agent. Each memory remembers all of the associations around all of its entity mentions, across all sources.

SaffronEnterprise™ allows a massive number of such agents to collaborate. Analogous to information sharing among human agents, these memories enable enhanced decision- making through the discovery of explicit and implicit links between entities. Each memory can associate millions of attributes with each other and there can be millions of Saffron memories. This complexity, millions of attributes in millions of memories, is what sets Saffron apart – the ability to handle context at scale.

Scalable Solution

SaffronEnterprise succeeds where other solutions fail due to its patented, memory-based technology. It is capable of building and maintaining extremely large associative memories that allow applications to automatically discover, store and retrieve relationships between entities. Each memory can associate millions of entity attributes and SaffronEnterprise can scale to millions of associative memories and beyond. These associative memories mimic the processes of the most intelligent machine on earth – the human brain – enabling SaffronEnterprise to provide unprecedented scalability and response times. Unlike relational databases that use a limited number of primary keys for index-based search, associative memories generated by SaffronEnterprise gather and store all entity associations. This allows the direct lookup of entity associations without requiring complex, inefficient and unscalable relational joins.

Timely Information

SaffronEnterprise assimilates enterprise data and remembers every entity characteristic at the granular level. As an organisation acquires new content, SaffronEnterprise incrementally and dynamically updates its network of interrelated entities. Existing entities have their memories updated. New entities are created, added to the network as nodes and virtually linked with other entities based on related information in documents and other datasources. This new knowledge can now be immediately accessed for analysis without requiring a lengthy batch ingestion or update process.

Context-Driven

SaffronEnterprise automatically learns the context and conditions that link entities across disparate data sources. This contextual understanding is vital as entities can be associated in many different ways. By understanding how entities are associated, SaffronEnterprise can consider non-obvious relationships and cut through noisy data to facilitate entity analysis. While all entity-to-entity associations are preserved, SaffronEnterprise enables applications to express the links most relevant to the task at hand. This context-driven approach provides enhanced decision making through the discovery of explicit and implicit entity associations. SaffronEnterprise delivers a comprehensive, distributed service platform that creates and manages dynamic entity networks. It powers existing applications that require detailed knowledge of entity information and associations. SaffronEnterprise can also be used to develop new applications requiring large-scale entity association mining. No matter the target application, SaffronEnterprise is an essential component of any cutting-edge entity analytics solution.