From March 14 to March 18, 2016 the  10th International Conference on Language and Automata Theory and Applications (LATA 20161) was held in Prague. It was organized by the Department of Theoretical Computer Science of Czech Technical University in Prague and the Research Group on Mathematical Linguistics of Rovira i Virgili University.

Several presentations included research related to the analysis of media resources such as images, text or metadata. From a MICO perspective, the following talks appeared to be most interesting:

– In an invited talk entitled “Automata for Ontologies” Frank Wolter discussed how automata theory can be used to analyze different reasoning problems for description logic ontologies. One example for such an ontology is the initiative by Bing, Google and Yahoo! which proposed a standard to enrich website content with metadata.

– In the talk “Colored Nested Words” Dana Fisman introduced the model of colored nested words which allows hierarchical structures that are not properly nested. This enables for instance the processing of XML-documents containing mismatches such as missing close-tags.

– The talk “Inferring a Relax NG Schema from XML Documents” dealt with the problem that in practice many XML documents come without a valid or with an incorrect schema. Guen-Hae Kim, Sang-Ki Ko, and Yo-Sub Han developed therefore an inference system that generates a Relax NG schema from a set of XML documents.

– In ”Reasoning with Prototypes in the Description Logic ALC Using Weighted
Tree Automata” Franz Baader and Andreas Ecke introduced a way to use prototypes in description logics, thus allowing for a better way of handling concepts whose definition contains typical properties which individual instances may lack. (E.g., cups usually have one handle, but some atypical ones may have two or none.)

From Umeå University, research partner of MICO, Johannes Blum and Petter Ericson presented the ongoing research on DAG languages and grammars, that can be used to deal with meaning representations in natural language processing. In the talk “Properties of Regular DAG Languages” a new automata model was introduced that describes regular DAG languages whereas in the talk “Between a Rock and a Hard Place – Uniform Parsing for Hyperedge Replacement DAG Grammars” a low degree polynomial-time algorithm for a restricted class of hyperedge-replacement grammars was presented.

For further information the proceedings are available as LNCS volume 9618

Johannes Blum, Umeå, 2016-03-23