Our breakthrough
Our products target allowing the automatic interpretation of unstructured and structured data. At the center of our products is the ability to cross the semantic gap between users and machines, by allowing the construction of machines with embedded commonsense knowledge. This supports the construction of flexible systems which are able to interpret language and its variations.
At the center of our products is the use of distributional semantic models, which enable the extraction of large amounts of commonsense knowledge from unstructured text.
Products
Databases are growing in size and complexity. StarGraph is an open source NoSQL platform for building and interacting with large and sparse knowledge graphs. StarGraph supports users in the interaction and knowledge discovery over knowledge graphs. By leveraging on large amounts of commonsense knowledge extracted from the Web, StarGraph is a database with built-in semantics, supporting an easier user interaction between users and data. Key features include:
- Query by meaning: Users do not need to understand the schema behind the database in order to query the data. In this case, StarGraph matches the meaning of the query to the meaning of the data.
- Built-in entity recognition and linking: Integrate structured and unstructured data seamlessly. StarGraph supports better data exploration and knowledge discovery by allowing the integration of unstructured and structured data.
- Knowledge discovery: Use the large amounts of commonsense data embedded in the database for discovering patterns in the data.
Core Features:
- NoSQL graph database.
- Semantic keyword search interface.
- Built-in semantic approximation.
- Scalable query execution performance.
- Supports data in RDF and JSON.
Natural is a Question Answering system which allows questions over knowledge graphs and text collections. Using StarGraph as a knowledge representation engine, Natural is a question answering system with a large amounts of embedded commonsense knowledge, extracted from the Web. This allows Natural to provide a robust interpretation of the meaning of the user queries and of the data.
Natural can be used for:
- Building customized question answering systems over organizational databases and texts.
- Building intelligent natural language interfaces for terminals, robots, and Web portals.
- Supporting domain experts on data exploration and knowledge discovery tasks.
For more information and for demonstrations please contact us.