We present a machine learning algorithm for refining the structure of a stochastic context–free grammar (SCFG). This algorithm consists of a heuristic for identifying structural errors and an operator for fixing them. The heuristic identifies nonterminals in the model SCFG that appear to be performing the function of two or more nonterminals in the target SCFG, and the operator attempts to rect...

This paper describes Grammar Learning by Partition Search, a general method for automatically constructing grammars for a range of parsing tasks. Given a base grammar, a training corpus, and a parsing task, Partition Search constructs an optimised probabilistic context-free grammar by searching a space of nonterminal set partitions, looking for a partition that maximises parsing performance and...

There are several interesting observations to be made concerning the tape complexity of context-free languages. An early result given by Lewis et al. (1965) is that every context-free language can be recognized by an off-line deterministic Turing machine of O((log n) 2) tape complexity. This is still the best result known. Sudborough (1975) shows that if all linear context-free languages can be...

Best-rst probabilistic chart parsing attempts to parse eeciently by working on edges that are judged \best" by some probabilistic gure of merit (FOM). Recent work has used probabilistic context-free grammars (PCFGs) to assign probabilities to constituents, and to use these probabilities as the starting point for the FOM. This paper extends this approach to using a probabilistic FOM to judge edg...

Pseudoknots are a frequent RNA structure that assumes essential roles for varied biocatalyst cell’s functions. One of the most challenging fields in bioinformatics is the prediction of this secondary structure based on the base-pair sequence that dictates it. Previously, a model adapted from computational linguistics – Stochastic Context Free Grammars (SCFG) – has been used to predict RNA secon...

This paper describes an application of Grammar Learning by Partition Search to noun phrase extraction, an essential task in information extraction and many other NLP applications. Grammar Learning by Partition Search is a general method for automatically constructing grammars for a range of parsing tasks; it constructs an optimised probabilistic context-free grammar by searching a space of nont...

This paper describes an application of Grammar Learning by Partition Search to noun phrase extraction, an essential task in information extraction and many other NLP applications. Grammar Learning by Partition Search is a general method for automatically constructing grammars for a range of parsing tasks; it constructs an optimised probabilistic context-free grammar by searching a space of nont...

We describe an extension of Earley's parser for stochastic context-free grammars that computes the following quantities given a stochastic context-free grammar and an input string: a) probabilities of successive prefixes being generated by the grammar; b) probabilities of substrings being generated by the nonterminals, including the entire string being generated by the grammar; c) most likely (...

The introduction of the adjoining operation to context-free grammars comes at high costs: The worst case time complexity of (Earley, 1968) is O n 3 ¡ , whereas Tree Adjoining Grammars have O n 6 ¡ ((Schabes, 1990)). Thus, avoiding adjoining as far as possible seems to be a good idea for reducing costs (e.g.) address this problem more radically by restricting the adjoining operation of TAGs such...

This paper describes Grammar Learning by Partition Search, a general method for automatically constructing grammars for a range of parsing tasks. Given a base grammar, a training corpus, and a parsing task, Partition Search constructs an optimised probabilistic context-free grammar by searching a space of nonterminal set partitions, looking for a partition that maximises parsing performance and...