SDC, Data Science and Knowledge

Knowledge and semantic technologies

From SDC, Data Science and Knowledge
Revision as of 12:08, 28 December 2015 by Zanni-merk (talk | contribs)
Jump to navigation Jump to search

The "Knowledge and Semantic Technologies" theme basically focuses on the design and implementation of formal models for the development of Knowledge Based Systems (KBS).

A KBS is software that reproduces the behavior of a human expert performing an intellectual task in a specific area. It is based on the explicit nature of knowledge, which is formalized in different ways. Among these formal models, ontologies are formalized and structured representations of the vocabulary specific to a certain area of study. Ontologies are commonly used with a set of rules which are chained to simulate the reasoning of a human expert. This traditional architecture has drawbacks, associated, mainly, with the difficulties that appear during the knowledge elicitation process from experts; and also with the noncompleteness of the formal conceptual model obtained after the elicitation.

In fact, as the knowledge base can be incomplete, there could be problems that this traditional architecture cannot solve. Reasoning and analysis of this incomplete knowledge implies that it is needed to take advantage of the experience acquired from the interventions of human experts when the traditional system does not lead to satisfactory results;

The originality of our works is founded on the proposal of the modula architecture KREM (Knowledge, Rules, Experience and Meta-Knowledge) to deal with the aforementined drawbacks. to incorporate the capitalization of experience with the goal of improving decision-making.

Because to be effective, decision-making must result from reasoning and analysis of domain knowledge, also taking into account the experience and expertise of decision-makers. As a consequence, it is needed to capitalize them to take advantage of the experience acquired from the interventions of human experts when the traditional system does not lead to satisfactory results.

The use of meta-knowledge to steer the execution of the whole system is also necessary. Meta-knowledge is knowledge about domain knowledge, about rules or about experience. This metaknowledge can take the form of context, culture or protocols to use this knowledge. Context is information that characterizes a situation in relation to interaction among human-beings, applications and their environment, and can be of four types: identity, place, status or time. Culture metaknowledge tries to take into account the fact that decisions are made differently depending on the country or culture. And finally protocols And finally protocols may include strategies or problem-solving heuristics for the task to be done (for example, in the case of medical diagnosis, the protocols used by physicians change according to the type of symptoms or the suspected illness).

Therefore, the proposed components of the architecture are:

  • The Knowledge component that contains the domain knowledge to operate, by means of different domain ontologies to be developed.
  • The Rules component that allows different types of reasoning (monotone, spatial, temporal, fuzzy, or other) depending on the application.
  • The Experience component that allows the capitalization and reuse of prior knowledge.
  • The Meta-knowledge component, including knowledge about the other three bricks that depends on the problem.

The way the domain knowledge is formalized will shape the way the rules are expressed. Experience will come complete the available knowledge and rules. Finally, meta-knowledge will directly interact with the rules and the experience to indicate which rules (coming from experience or from the initial rule set) need to be launched according to the context of the problem to solve.

The main areas of application of these tools are:

  • Analysis of Remote Sensing Images
  • Smart Factories and Processes
  • Environment and Sustainable Development
  • Steering of Complex Biomolecular Networks