Products & Services

  • Eliminate data preprocessing steps and enable Analysts to focus on analysis.

  • Data Packaging Service collects data from public or private data sources, transforms the data for integration with other sources, and delivers the results as self-contained files.

DeciSym AI Engine

Eliminate data preprocessing steps and enable Analysts to focus on analysis.

  • See PDF.

  • Analysts frequently expend time and effort identifying data sources, collecting them for analysis, bulk loading them into an analysis database or tool, normalizing/transforming them into a canonical or interoperable format, and indexing the contents. Data freshness characteristics may require the entire pipeline to be redone for subsequent analysis. The pipeline is typically appropriate only for one group or class of analysis and cannot be reused organization-wide. DeciSym Data Packaging Service enables organizations to shift the effort of data source collection, normalization, and indexing away from internal staff to a service provider. DeciSym AI Engine is a triple-store database that enables Analysts to immediately query Data Packages, skipping the cost and effort of pre-processing.

  • Custom interoperable Data Fabrics can be created at any organizational level, from air-gapped enclaves to globally distributed and networked sites. The DeciSym AI Engine can perform queries that can cross local storage, the collection of copies from hosted storage lacking compute (e.g., AWS S3, SCP, etc.), and remotely hosted instances having both storage and remote compute (e.g., dedicated servers, VMs, etc.).

  • Ontologies provide the mechanism for systematic, distributed data integration. Traditional Relational Database Management Systems (RDBMS) require data structures to be predefined before data can be included in centralized data warehouses or data lakes. Ontologies embrace distributed data with heterogeneity, allowing different communities of practice to use their own vocabulary while still integrating with data from other communities.

  • The DeciSym AI Engine is implemented in Rust on top of foundational open-source, standards compliant triple stores and RDF processing libraries. It leverages our partner Sylabs' advanced encryption, access control, and compression technology, ensuring data security at all processing stages. The Engine is deployed as a simple CLI with customization points, such as feeding organization-specific dashboards.

DeciSym Data Packaging Service

Data Packaging Service collects data from public or private data sources, transforms the data for integration with other sources, and delivers the results as self-contained files.

    • DeciSym transforms the data into a Resource Description Framework (RDF) encoding.

    • The RDF is aligned to interoperable ontologies that comply with the Web Ontology Language (OWL) standard.

  • By leveraging our partner Sylabs' advanced encryption, access control, and compression technology, DeciSym’s Data Packaging Service ensures a reliable and secure data supply chain for analytics, AI/ML, and HPC workflows across distributed and disparate environments.

    • Compared to alternative Extract, Transform, and Load (ETL) services, the Data Packaging Services transform into a standardized, canonical RDF encoding. The transformed data is aligned to an ontology to support data integration and query across multiple sources. The ontology enables inference operations.

    • Data Packages include a complete index to respond directly to SPARQL queries, eliminating the need to load the data into another system.

    • Our approach reduces the data handling, preprocessing, and other touchpoints that slow down and distract Data Analysts

    • The data packages are well-suited for High-Performance Computing (HPC) infrastructure. Data is quickly distributed to individual compute nodes without the centralized services/daemons required by traditional databases. DeciSym’s approach results in better HPC compute node utilization.

    • Data Packages are also highly compressed and deployable to Edge computing nodes with limited storage. The self-contained index, compression, and encryption combine to enable distributed computation architectures and deployments from Edge-to-Exascale.