Data Fabric for Machine Learning

Digital Transformation is about connecting data with the people that use it. But modern systems are a complex mix of cloud applications, streaming data and traditional databases. Ops teams are under constant pressure to provide faster access to information from increasingly diverse sources. StreamScape's platform is a data fabric powered by cognitive AI that makes it easy for organizations of all size to identify and gather critical data, accelerating delivery and empowering cross-functional teams.

Data Fabrics are a new way to manage and integrate data, that overcome the limits of prior technologies. They address the long-standing challenges faced by operations teams, system integrators and analytics leaders. As the data landscape grows, so does the need to aquire accurate and relevant, high-value information. StreamScape lets you cut through the noise and focus on building modern, AI powered applications that leverage machine learning, cloud, web and big data storage. Use the data fabric to personalize user experience, understand trends, automate research tasks and augment human intelligence with machine-guided decisions.

Multi-Model Data Store

The data fabric uses Application Dataspaces™ to store and query information in structured or unstructured format. Data collections may be in-memory, persistent or hybrid based on application needs. This multi-model approach lets users easily add a rich semantic layer to their data, providing deeper meaning and context without the complexity of schema changes.

User defined Semantic Types make it easy to work with data that describes real world things like Medical Records, Insurance Claims or Financial Instruments, allowing users to bridge the gap between business expectations and the challenges of Machine Learning. Combining in-memory speed with parallel processing and data compression lets the fabric perform complex, multi-dimensional analysis 100-1,000 times faster than other solutions at a fraction of the cost.

Semagraphs, Facets + Aspects

Facets and Aspects are specialized AI types used to describe (annotate) dataspace content, and may be added to any data element. The annotation attributes let users infer relationships and discover connections between data elements, presenting results as a table, document or a graph.

Semagraphs are vector indexes created by the platform's language processing AI. They can be built from any data source and used to automatically annotate dataspace contents based on semantic similarity of words and phrases. Annotated data collections can be joined by inference instead of simple equality, allowing users to easily blend structured and unstructured data with results from fabric Text Search, Language Processing Models, Sentiment or Thematic Analysis.

Knowledge Graphs are the essential value creating components of a Data Fabric
- Gartner Group

Streaming Analytics + Traditional BI

All the tools you need, in one tool box! StreamScape's Event Fabric™ is the glue that holds everything together, allowing users to work with streaming data, relational tables, NoSQL, documents and Machine Learning models using a common SQL-based query language. Integrated stream analytics let you process large volumes of data efficiently using minimal hardware. A logical data layer makes it easy to combine data from many sources for easy access by cloud, mobile, desktop apps and visualization tools like Tableau or Power BI.

Data Virtualization

Access and query files, SQL Databases, No-SQL, Big Data and Cloud Storage content anywhere using Data Virtualization Services. Easily integrate data from disparate sources or load it into the dataspace. Call web API's or connect to Messaging systems like JMS or Kafka to process messages in real-time. Turn any data into an event stream that drives business process and Machine Learning pipelines.

Knowledge Graphs

A knowledge graph, also known as a semantic network, represents a model of real-world things such as places, people, events, business documents, or concepts—and illustrates the relationships between them. Links between graph entities may be defined as specific types, allowing users to query data by their relationships instead of matching values.

The data fabric stores knowledge graphs in Knowledge Set collections similar to a Triplestore. Unlike graph databases that require entity relationships to be specified upon data creation, StreamScape allows relationships to be setup dynamically based on triggers and actors, or via background tasks performed by the cognitive AI. Relationships can be direct or based on Semagraph lookups, Text Search, Classification models, or Inference Query offering a variety of powerful techniques for graph building.

Semantic Links

Define relationships between data collection elements using semantic links, allowing the data fabric to connect entities that are correlated or similar in concept. Links make it possible to build and refine knowledge graphs thru data exploration, classification or automated discovery, and construct Concept Spaces that represent likely relationships between data elements. This allows knowledge graphs to be materialized on-demand in response to queries or application needs and lets the data fabric operationalize graph creation, update and maintenance.