|
SOLUTIONS
Blue Reference uses the Paradigm Platform to deliver search-based applications to help scientists and engineers to optimize business performance by connecting information, science objects and people across research and development networks. Search-based applications combine search, science object management and content analytics with collaborative technologies and workflow. Search-based applications employ tailored user interfaces designed for specific tasks or workflows. Illustrative search-based applications include the following:
Paradigm Platform for Scientific Search-Based Applications
The Paradigm Platform is an integrated collection of software services designed for construction of search-based applications involving science objects. The Paradigm Platform is constructed as a Service Oriented Architecture built on top of the Microsoft .NET 4.0 Framework.
The Paradigm Platform is written in native C# for 32- and 64 bit environments, has a documented API, and can be extended through the Paradigm SDK. The Paradigm Platform is comprised of three principal service layers offering a broad array of capabilities including the following:
Search and Connection Services
- Unified information access across heterogeneous structured and unstructured data
- Designed for non-expert searchers; conversational queries; no knowledge of schema required
- Multiple query type support: natural language, Boolean, semantic, fuzzy and science objects
- Relevance-ranked results with selectable relevance scoring and ranking methods
- Metadata services for display and management of metadata and controlled vocabulary
- Dynamic connector services for connection to popular data sources like File Shares, SharePoint, EMC Documentum, Nugenesis SDMS, Accelrys ELN and major LIMS systems
- Fine grained control of security across multiple sources
- Choice of client deployments: in browser, WPF or Silverlight application, or as custom embedded .NET component in third-party application
Science Object Services
- Comprehensive support for abstracting scientific content as science objects used in cheminformatics and bioinformatics
- Full support for sequences (DNA, RNA, AA) as biology objects including reading/writing, import/export, searching (exact, subsequence), ranking (similarity, alignment), display, property/descriptor and fingerprint calculation, and metadata extraction
- Full support for structures (connection tables) and reactions as chemistry objects including reading/writing, import/export, searching (exact, substructure, transformation), ranking (similarity), display, property/descriptor and fingerprint calculation, and metadata extraction
- Support for analytical data as chromatography and spectral objects including reading/writing, import/export, searching (exact), ranking (similarity), display, property/descriptor and fingerprint calculation, and metadata extraction
- Support for biological images as image objects including reading/writing, import/export, searching (exact), ranking (similarity), display, property/descriptor and fingerprint calculation, and metadata extraction
- Support for structured data as dataset objects including reading/writing, import/export, searching (exact), ranking (similarity), display, calculation, and metadata extraction
Content Analytics Services
- Automated methods to read, recognize, understand and extract valuable metadata from unstructured content via a content analytics processing pipeline
- Content analytics processing pipeline operates on content whether as digital text, text as image (via Optical Character Recognition), or text in images
- Content analytics processing pipeline for core metadata extraction including titles, contributors/authors, affiliations, addresses and abstracts
- Recognize and extract science entities from content. Map science entities to science objects and normalize. Science objects include biology, chemistry, chromatography, spectral, image and datasets
- Content analytics processing pipeline for chemical name (text) entity extraction and name to structure conversion
- Content analytics processing pipeline for chemical image (drawing) entity extraction (via Optical Structure Recognition) and name to structure conversion
- SDK available for building you own custom processing pipeline using state-of-the-art machine learning technology
|
