Learning linear grammars from structural information
نویسندگان
چکیده
A bstract.Lin ear lan guage class is a su bclass of context-free lan guage class. In t his paper, we propose an algorithm to learn linear languages from structural information of t heir strings. We compare our algorithm with other ad apt ed algorit hm from Radhakrishnan an d Nagaraja [RN1]. The proposed method and the adapt ed algorithm are heuristic techniques for the learnin g t asks, and they are useful when only positive structural data is availab le.
منابع مشابه
Learning context-free grammars from stochastic structural information
We consider the problem of learning context-free grammars from stochastic structural data. For this purpose, we have developed an algorithm (tlips) which identiies any rational tree set from stochastic samples and approximates the probability distribution of the trees in the language. The procedure identiies equivalent subtrees in the sample and outputs the hypothesis in linear time with the nu...
متن کاملLearning Context-Sensitive Languages from Linear Structural Information
In this work we propose a method to infer context-sensitive languages from positive structural examples produced by linear grammars. Our approach is based on a representation theorem induced by two operations over strings: duplication and reversal. The inference method produces an acceptor device which is an unconventional model of computation based on biomolecules (DNA computing). We prove tha...
متن کاملTranslation as Linear Transduction Models and Algorithms for Efficient Learning in Statistical Machine Translation
Saers, M. 2011. Translation as Linear Transduction. Models and Algorithms for Efficient Learning in Statistical Machine Translation. Acta Universitatis Upsaliensis. Studia Linguistica Upsaliensia 9. 133 pp. Uppsala. ISBN 978-91-554-7976-3. Automatic translation has seen tremendous progress in recent years, mainly thanks to statistical methods applied to large parallel corpora. Transductions rep...
متن کاملLearning Equal Matrix Grammers and Multitape Automata with Structural Information
Equal matrix grammars are one of paralel rewriting systems. This type of parallel rewriting systems has been investigated in several areas such as L-systems and the syntactic pattern recognition. In this paper, we introduce two types of structural information and show that if these are available and there is a method of learning regular sets in polynomial time, then for any positive integer $k$...
متن کاملA Comparison of Rule Extraction for Different Recurrent Neural Network Models and Grammatical Complexity
It has been shown that rules can be extracted from highly non-linear, recursive models such as recurrent neural networks (RNNs). The RNN models mostly investigated include both Elman networks and second-order recurrent networks. Recently, new types of RNNs have demonstrated superior power in handling many machine learning tasks, especially when structural data is involved such as language model...
متن کامل