Intellisem is a company which focuses on the development of the next generation of intelligent and semantic information systems.

Our vision:

“Enabling natural communication with systems”

The Zeitgeist

The growth in the availability of data brings an enormous opportunity for better and more rational decision making. Organizations which are dependent on knowledge can use data to optimize their processes, address critical problems and innovate. However, accessing the value of the data currently depends on a time-consuming process of accessing the knowledge inside the data, which in many cases can be deeply dependent on the availability of large technical teams to support the querying, analysis and visualization of knowledge.

In many cases, even the simplest query operations are time consuming and depend on the availability of IT professionals, creating a cycle in which the full value of data is not accessed and becomes, in many contexts, economically unfeasible.

We address a real need:

The demand to access large amounts of heterogeneous structured data is emerging as a trend for many users and applications.

And what does this mean?

However, the effort involved in querying heterogeneous and distributed third-party databases can create major barriers for data consumers.

What is the problem?

At the core of this problem is the semantic gap between the way users express their information needs and the representation of the data.

What we offer?

At IntelliSem we have developed StarGraph and Natural, a robust framework to allow users to interact with large-scale and sparse databases using natural language queries. It is a tool which works as a “Google” for structured data, in which users can search and query sparse structured data.

a

Users can access information with queries such as:

• “Give me all schools with the worst performance indicators in the Munich area?”
• “How many road construction sites were there in Passau on the period of November 2015?”
• “Which were the construction contractors and the value of their contracts employed by the State Government of Bavaria?”

What does Natural do?

Natural provides a natural language interface and an associated semantic index to support an increased level of vocabulary independence for queries over Linked Data/Semantic Web datasets, using a distributional-compositional semantics approach.

Distributional semantics – what is this?

Distributional semantics focuses on the automatic construction of a semantic model based on the statistical distribution of co-occurring words in large-scale texts.

a

The main center of gravity for our technologies relates to knowledge discovery and is inspired by research in distributional semantics, an area that may be regarded as one of the most strategic research areas in semantic computing for the next years.

a

One of the key areas of our research relates to the critical technology for big data analytics in the form of distributional semantics. Distributional semantics can be regarded as the new semantics for the era of big data. Distributional semantics is a promising field of research that can contribute to sense-making out of complex data artifacts which may come from several sources such as text, structured data, images, sounds and videos in order to help adapt services to specific user groups and personal and environmental contexts.

a

Semantic models we employ target features like:

a

• principled semantic approximation approach with low adaptation effort (independent from manually created resources such as ontologies, thesauri or dictionaries);
• comprehensive semantic matching supported by the inclusion of large volumes of distributional (unstructured) commonsense knowledge into the semantic approximation process;
• expressive natural language queries.

a

It is the last part that concentrates much interest from our pilot users and customers (how can people express questions in natural language and get meaningful and preferably correct) answers, however all three above mentioned features are important.

The use of distributional semantic models also supports the creation of a solution which can be easily transported to other languages. Additionally, users can input new information into the knowledge graph using natural language. The proposed model build upon one decade of research by the company partners on question answering, semantic search, natural language queries, distributional semantics and information extraction. The existing platform, StarGraph, targets reducing the barriers for accessing and creating data using natural language as the core data interaction and data creation medium.

And how does the market move?

Google Knowledge Graph is only a recent example of the benefits of enabling the use of large-scale structured data resources may bring to applications.

Additionally, during the last years, Open Linked Data emerged as a standard for publishing structured data on the Web, playing a fundamental role in enabling the next generation of applications driven by rich Web data. However, the effort involved in querying heterogeneous and distributed third-party Linked Data sources on the Web creates barriers for data consumers.

And again: what does this mean?

In order to query datasets, users need to discover the datasets of interest, understand the structure and vocabularies used in these datasets, and then finally formulate the query using the syntax of a structured query language (such as SPARQL or SQL).

Ideally users should be able to express their information needs without being aware of the dataset vocabulary (or ‘schema’), delegating the query formulation process to a query engine.

In fact, most semantic models have dealt with particular types of constructions, and have been carried out under very simplifying assumptions, in true lab conditions.

If these idealizations are removed it is not clear at all that modern semantics can give a full account of all but the simplest models/statements.

a

Actionable data

At Intellisem we provide our customers with actionable data – a resource that brings the potential of increasing rationalization levels in the management of corporate resources, where decisions can be justified by a principled quantitative analysis. The use of  actionable data as a mediator for decision making and for creating a dialogue along the different corporate levels and streams, brings the opportunity of increasing the quality of managing corporate resources and safeguarding your shareholders’ interests.

Moreover, with easier and broader access to actionable data, companies can better understand the market demands and meet individual customer needs.