The Omphalos Context-Free Grammar Learning Competition

Starkie, Bradford; Coste, Fran » cois; Zaanen, Menno van

doi:10.1007/978-3-540-30195-0_3

Cited by 18 publications

(18 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We selected a number of languages that have been proposed in the literature, generally from small alphabets. For comparison, we also evaluated it on one of the more complex grammars from the Omphalos context free grammatical inference competition (Starkie et al, 2004); this is at the state of the art for context free grammatical inference.…”

Section: Methodsmentioning

confidence: 99%

“…Negative data is more of a problem. Simply generating random strings in a similar way does not produce a test set sufficiently difficult to distinguish the true hypothesis from a similar but incorrect one, without using astronomical amounts of data -the same problem was encountered by the organisers of the Omphalos competition (Starkie et al, 2004). We thus generated the negative data sets 50% from random strings from a uniform distribution over strings, and 50% drawn from languages that are close to the true one.…”

Section: Methodsmentioning

confidence: 99%

“…For context free languages, some recent approaches have had some limited success (Starkie et al, 2004). However, it is well known that certain features of natural language cannot be described by context free grammars, and require the power of mildly context sensitive grammars, which cannot be learned with current techniques.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Languages as Hyperplanes: Grammatical Inference with String Kernels

Clark

Florêncio

Watkins

2006

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Using string kernels, languages can be represented as hyperplanes in a high dimensional feature space. We present a new family of grammatical inference algorithms based on this idea. We demonstrate that some mildly context sensitive languages can be represented in this way and it is possible to efficiently learn these using kernel PCA. We present some experiments demonstrating the effectiveness of this approach on some standard examples of context sensitive languages using small synthetic data sets.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Languages as Hyperplanes: Grammatical Inference with String Kernels

Clark

Florêncio

Watkins

2006

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…The first case includes well known benchmarks based on the ABBADINGO or OMPHALOS competitions Starkie, Coste, & van Zaanen, 2004). Other large benchmarks have been made available by Oliveira & Silva (2001) or the GOWACHIN system (Lang, Pearlmutter, & Coste, 1998).…”

Section: About Grammatical Inference Experimentsmentioning

confidence: 99%

“…In 2003 there has been renewed interest in the topic (de la Higuera et al, 2003): the OMPHALOS context-free language learning competition (Starkie, Coste, & van Zaanen, 2004) was launched, where state of the art techniques were unable to solve even the easiest tasks. The method (Clark, 2006) that obtained best results used a variety of information about the distributions of the symbols, substitution graphs and context.…”

Section: Introductionmentioning

confidence: 99%

LARS: A learning algorithm for rewriting systems

2006

View full text Add to dashboard Cite

Whereas there is a number of methods and algorithms to learn regular languages, moving up the Chomsky hierarchy is proving to be a challenging task. Indeed, several theoretical barriers make the class of context-free languages hard to learn. To tackle these barriers, we choose to change the way we represent these languages. Among the formalisms that allow the definition of classes of languages, the one of string-rewriting systems (SRS) has outstanding properties. We introduce a new type of SRS's, called Delimited SRS (DSRS), that are expressive enough to define, in a uniform way, a noteworthy and non trivial class of languages that contains all the regular languages, {a n b n : n ≥ 0}, {w ∈ {a, b} * : |w| a = |w| b }, the parenthesis languages of Dyck, the language of Lukasiewicz, and many others. Moreover, DSRS's constitute an efficient (often linear) parsing device for strings, and are thus promising candidates in forthcoming applications of grammatical inference. In this paper, we pioneer the problem of their learnability. We propose a novel and sound algorithm (called LARS) which identifies a large subclass of them in polynomial time (but not data). We illustrate the execution of our algorithm through several examples, discuss the position of the class in the Chomsky hierarchy and finally raise some open questions and research directions.

show abstract