Adrian Horia Dediu

Special Issue on Second International Conference on the Theory and Practice of Natural Computing, TPNC 2013

This special issue of the journal ‘Soft Computing—AFusion of Foundations, Methodologies and Applications’ offers extended versions of some of the best papers presented at the Second International Conference on the Theory and Practice of Natural Computing, TPNC 2013, held in Caceres, Spain, on December 3–5, 2013, under the organisation of the Research Group on Mathematical Linguistics (GRLMC) from Rovira i Virgili University in Tarragona, Spain and the Computer Architecture and Logic Design Group (ARCO) from University of Extremadura, Spain. TPNC 2013 was the second event in a series dedicated to host and promote research in a wide spectrum of computational models, methods and techniques inspired by information processing in nature. We encouraged both theoretical contributions in soft computing, computing architectures, and formal models as well as solutions for practical problems based on natural computing methods. Out of 47 submissions to the conference, 18 papers were accepted (which represents a competitive acceptance rate of about 38 %). Among them, the authors of 13 papers were invited to submit to this special issue. Each submission was reviewed by at least two experts and, on the basis of their comments, the guest editors decided to accept six papers for this special issue (which represents an acceptance rate of about 13 % out of the submissions to the conference).

Learning DFA from correction and equivalence queries

In active learning, membership queries and equivalence que- ries have established themselves as the standard combination to be used. However, they are quite “unnatural” for real learning environments (membership queries are oversimplified and equivalence queries do not have a correspondence in a real life setting). Based on several linguistic arguments that support the presence of corrections in childrens language acquisition, we propose another kind of query called correction query. We provide an algorithm that learns DFA using correction and equivalence queries in polynomial time. Despite the fact that the worst case complexity of our algorithm is not better than Angluins algorithm, we show through a large number of experiments that the average number of queries is considerably reduced by using correction queries.

Simulating evolutionary algorithms with eco-grammar systems

Starting from an arbitrary evolutionary algorithm, we construct an eco-grammar system that simulates the EAs behavior. There are only several practical applications of eco-grammar systems and our approach tries to bring a new light in this field. We believe that our research opens a new perspective also for evolutionary algorithms that can benefit from the theoretical results obtained in the framework of eco-grammar systems.

Contextual hypergraph grammars – a new approach to the generation of hypergraph languages

In this paper, we introduce contextual hypergraph grammars, which generalize the total contextual string grammars. We study the position of the class of languages generated by contextual hypergraph grammars in comparison with graph languages generated by hyperedge replacement grammars and double-pushout hypergraph grammars. Moreover, several examples show the potential of the new class of grammars.

Eco-Grammar systems as models for parallel evolutionary algorithms

Evolutionary Algorithms (EAs), biological inspired searching techniques, represent a research domain where theoretical proofs are still missing. Due to the lack of theoretical foundations, an extensive experimental work developed many variations of the basic model. Remarkable tendencies such as variable control parameters or parallel populations try to overcome the stagnation observed at the end of evolutions. We tried to study from theoretical point of view the possibility of modelling parallel EAs using Eco-grammar systems. We expect that our research opens a new perspective over EAs behavior and our framework can bring theoretical results that will lead to new recommendations for EAs architectures as well as for specific details requested by practical problems.

Trees as contexts in formal language generation

Initially designed to generate languages without rewriting of some nonterminals, contextual grammars are used in formal language theories as well as models for several particular aspects of natural languages. However, despite their power, contextual grammars do not provide a structural description of the generated languages. We present several modalities to add structures to classical contextual grammars. We are studying the relations of these structured contextual languages comparing with other structured languages. Several examples show the potential of the newly introduced grammars.