Data representation and reasoning
Data representation and reasoning is the arena of (AI) artificial intelligence. It is dedicated to representing information around the world. That information is in a form that a computer system may utilize to solve complex tasks. Those tasks are for example identifying a medical condition or having a dialog in a natural language. Data representation joins findings from psychology about how humans solve problems and represent knowledge in order to design formalisms. That would make complex systems easier to design and build. Data representation and reasoning moreover include findings from logic to automate many kinds of reasoning, for instance, the application of rules or the dealings of sets and subsets.
Data representation does focus on designing computer representations. The explanation for knowledge representation is that conventional procedural code is not the best formalism to use to solve difficult problems. Data representation makes difficult software stress-free to define and keep than procedural code and may be used in expert systems. For instance, speaking to experts in terms of business rules somewhat than code reduces the semantic gap between users and developers. This makes the development of complex systems more practical. Data representation drives hand in hand with automated reasoning. Because one of the key purposes of explicitly representing knowledge is to be capable of reason near knowledge. This is to make implications, declare new knowledge, etc. Practically all data representation languages have a reasoning or inference engine as part of the system.
The main trade-off in the design of knowledge representation formalism is that amid clarity and practicality. The First Order Logic (FOL) is the last knowledge representation formalism in terms of expressive power and compactness. Though, FOL has two disadvantages by way of knowledge representation formalism. First-order logic may be threatening even for several software developers.
Languages that do not have the whole official power of FOL may still deliver close to the similar expressive power with a user interface. That is additional practice for the average developer to understand. The subject of the practicality of implementation is that FOL in some ways is also expressive. This is possible through FOL to create statements that would reason a system to never dismiss if it tried to verify them.
Therefore, a subset of FOL might be together easier to use and more practical to implement. This was a dynamic motivation behind rule-based expert systems.
Five different roles to explore a Data representation framework
1. Data representation is the finest basically a replacement. This is a substitute for the thing itself. It is used to allow an entity to control consequences by thinking somewhat than acting, for example, by reasoning near the world rather than taking action in it.
2. This is a set of ontological commitments. These are a response to the question. For example, in what terms must I think about the world?
3. It is an incomplete theory of intelligent reasoning. It is said in terms of three mechanisms:
(i) The representation’s important the concept of intelligent reasoning
(ii) The set of implications the representation sanctions
(iii) The set of inferences it endorses
4. This is a medium for pragmatically efficient computation, for example, the computational environment in which thinking is able. One role to this pragmatic efficiency is delivered by the guidance a representation provides for organizing information. This is so as to assist in making the recommended inferences.
5. This is a medium of human appearance, for instance, a language in which we say things about the world.
Semantic Web, knowledge representation and reasoning
Data representation and reasoning are the main permitting technology for the Semantic Web. Languages built on the Frame model by automatic classification deliver a layer of semantics on top of the current Internet. It would be likely to define logical queries and find pages that map to those queries, relatively than searching through text strings as is typical today. The preset reasoning component in these systems is an engine. That engine is known as the classifier. Classifiers give emphasis on the subsumption relations in a knowledge base quite than rules. A classifier may conclude new classes and energetically change the ontology by way of new information becomes obtainable. This competence is perfect for the ever-changing and developing information space of the Internet.
The Semantic Web mixes concepts from knowledge representation and reasoning by means of markup languages based on XML. The Resource Description Framework (RDF) offers the basic capabilities to define knowledge-based objects on the Internet with simple features. The Web Ontology Language (OWL) enhances extra semantics and adds with automatic classification reasons.