Search

At the GU Dialogue Systems Lab in Göteborg we are embedding a conversational agent platform - the Current platform - in the Oz programming language. Current is based on a simple and intuitive characterization of conversational agents as interactive transducers, and on the fact that this characterization has a very direct implementation in Oz. Concurrency as offered by Oz allows our agents to ‘perceive', ‘think' and ‘act' at the same time. Concurrency in combination with streams allow our agents to process input in an incremental manner, even when the original underlying algorithms are batch-oriented. Concurrency and streams in combination with ports allow us to specify the ‘toplevel' transducer as a network of components - an interesting and highly modular architecture. We believe that software tools for specifying networks should have a strong visual aspect, and we have developed a ‘visual programming language' and an IDE to support it. Also, we have found that if we specify the non-visual aspects of transducers and other components as class definitions that inherit the methods responsible for the interpretation of condition-action rules, regular expressions, grammars, dialogue management scripts, etc. from (abstract) classes provided by separate modules, we are able to hide most of the gory details involving threads, streams and ports from the agent developer. [ABSTRACT FROM AUTHOR]

This paper presents logical reconstructions of four different methods for part of speech tagging: Finite State Intersection Grammar, HMM tagging, Brill tagging, and Constraint Grammar. Each reconstruction consists of a first-order logical theory and an inference relation that can be applied to the theory, in conjunction with a description of data, in order to solve the tagging problem. The reconstructed methods are compared along a number of dimensions including ontology, expressive power, mode of reasoning, uncertainty, underspecification, and robustness. It is argued that logical reconstruction of NLP methods in general can lead to a deeper understanding of the knowledge and reasoning involved, and of the ways in which different methods are related. [ABSTRACT FROM AUTHOR]