MICO contributes to the 21st International Conference on Implementation and Application of Automata (CIAA), and the co-located workshop on Trends in Tree Automata and Tree Transducers (TTATT) with two papers. The conference concerns research on all aspects of implementation and application of automata and related structures, including theoretical aspects, and is organised by Yo-Sub Han and his team at the Department of Computer Science of Yonsei University.
The first of the two MICO papers is written by J. Björklund, F. Drewes, and A. Jonsson and treats the extraction of the N best vertices in a weighted hyper graph, a theoretical problem with practical implications for several natural-language processing (NLP) tasks, for example,part-of-speech recognition and parsing. The second paper is written by S. Bensch, J. Björklund, and M. Kutrib and concerns so-called deterministic stack transducers, a computational model for reasoning about input-output system such as natural-language interfaces.
In particular CIAA is relevant for MICO, as many of the presentations concerned finite-state analysis and transformation of data. To mention a few examples, Calvo-Zaragoza et al. presented a technique for improved error-correction in automatic speech recognition; Woodman et al. analysed the matching time of regular expressions (REs), one of the most common tools in NLP; and Sulzman and Zhou Ming Lu used derivation-based diagnosis to understand RE ambiguity. More information about the events and the accepted papers is at the conference webpage.