Inference Technology & Architecture

Available as a web-based ASP hosting platform, Inference has combined advanced pattern-matching conceptual analytics and unlimited scalability with complete electronic discovery review and production tools. Because Inference is available from a standard internet browser without any  additional software or ActiveX controls, the platform is consistently available from anywhere, anytime, 24/7/365.

Search Technology

Inference utilizes a unique combination of Shannon’s Information Theory and Bayesian Inference to power its Analytics Suite. These algorithmic calculations, called Meaning Based Computing, extract patterns and identify similar trends, forming conceptual and contextual understanding of any piece of electronic data, whether structured, semi-structured or unstructured. The engine analyzes a system’s information and data, linking similar documents across the dataset, regardless of the type or format of the data, and in real time.

Because of this search methodology, Inference users can search across data types, languages, and records without the limitations associated with search technologies based on dictionary, thesaurus, or noun and noun phrase engines. Inference rapidly indexes volumes of data to enable search, retrieval and analytics, allowing the reviewer to determine relevance and responsiveness based on the actual meaning of the data. Therefore Inference’s powerful search methodology can be utilized during the early case assessment, meet-and-confer-prep, investigation, and full scale discovery and review phases of litigation to drill into data, cull down datasets, and organize review workflow to streamline and accelerate legal review.

Architecture & Scale

Inference contains no actual database. All data is held within a series of index engines, and all changes, updates, queries and analytics are performed inside those powerful engines. Inference uses multiple indices to support scale, and operations are guided, or distributed, through Query Parsers to send commands to individual engines allowing for consistent system performance.

Because of this architecture, Inference is completely scalable. Query speed can be maintained across the entire repository, and across simultaneous users. Additionally, Inference can support simultaneous, automatic updates and actions, while accepting rolling deliveries, without the need to re-index.