Abstract not found

. Attribute grammars can be considered as an extension of contextfree grammars, where the attributes are associated with grammar symbols, and the semantic rules define the values of the attributes. This formalism is widely applied for the specification and implementation of the compilation-oriented languages. The paper presents a method for learning semantic functions of attribute grammars which is a hard problem because semantic functions can also represent relations. The method uses background knowledge in learning semantic functions of S-attributed and L-attributed grammars. The given context-free grammar and the background knowledge allow one to restrict the space of relations and give a smaller representation of data. The basic idea of this method is that the learning problem of semantic functions is transformed to a propositional form and the hypothesis induced by a propositional learner is transformed back into semantic functions. Key words: attribute grammar, machine learning,...

This paper introduces a machinery called attributed top-down parsing automaton which performs top-down parsing of strings and, simultaneously, the evaluation of arbitrary noncircular attribute grammars. The strategy of the machinery is based on a single depth--first left-to-right traversal over the syntax tree. There is no need to traverse parts of the syntax tree more than once, and hence, the syntax tree itself does not have to be maintained. Attribute values are stored in a graph component, and values of attributes which are needed but not yet computed are represented by particular nodes. Values of attributes which refer to such uncomputed attributes are represented by trees over operation symbols in which pointers to the particular nodes at their leaves are maintained. Whenever eventually the needed attribute value is computed, it is glued into the graph at the appropriate nodes.

In this paper we present an application of the AGLEARN method to the part-of-speech (POS) tagging of Hungarian sentences. The task of the AGLEARN is to infer the semantic functions associated with the productions. In the learning process the grammar, background semantic functions and examples can be used. We applied the AGLEARN method to infer context rules to choose the correct tags. A corpus with about 100 000 pre-tagged words has been employed for training and testing. By using AGLEARN algorithm learning data sets are generated to the C 4.5 attribute value learner. These generated data contain information about the phrase structure of the sentences. A background attribute grammar has been used to determine the structural information. Our studies showed that using this structural background information the C 4.5 learner was able to infer more precise context rules. 1 1. Introduction Attribute grammars can be considered as an extension of context-free grammars, where the attributes...