The evaluation metric in generative grammar

PLOS ONE, 2018

Probabilistic proposals of Language of Thoughts (LoTs) can explain learning across different domains as statistical inference over a compositionally structured hypothesis space. While frameworks may differ on how a LoT may be implemented computationally, they all share the property that they are built from a set of atomic symbols and rules by which these symbols can be combined. In this work we propose an extra validation step for the set of atomic productions defined by the experimenter. It starts by expanding the defined LoT grammar for the cognitive domain with a broader set of arbitrary productions and then uses Bayesian inference to prune the productions from the experimental data. The result allows the researcher to validate that the resulting grammar still matches the intuitive grammar chosen for the domain. We then test this method in the language of geometry, a specific LoT model for geometrical sequence learning. Finally, despite the fact of the geometrical LoT not being a universal (i.e. Turing-complete) language, we show an empirical relation between a sequence's probability and its complexity consistent with the theoretical relationship for universal languages described by Levin's Coding Theorem.

Journal of Japanese Linguistics, 2010

This paper explores how the hypothetico-deductive method can be applied to research concerned with the properties of the language faculty. The paper first discusses how we can try to identify informant judgments that are likely a reflection of properties of the Computational System (or properties of the language faculty that are directly related to the Computational System), proposes a method of hypothesis testing in line with the hypothetico-deductive method, and provides an illustration by examining the predictions made under the lexical hypothesis that otagai in Japanese is a local anaphor.

Bucharest Working Papers in Linguistics XVII/2, 2015

Natural Language Engineering, 2003

Chapter 2 discussed the problem of quantifying the degree of resemblance between separate analyses of a given stretch of wording. We mentioned in 2.4 that the metric we use, the leaf-ancestor metric, is not the one standardly used in our discipline but, we believe, is considerably superior to the standard metric. In that chapter, the purpose for which we needed a tree-comparison metric was investigation of how far a set of parsing guidelines succeeded in predicting a unique analysis for a given language sample, so that separate human analysts dealing with the same samples would be constrained to come up with the same analyses.

This monograph (written in Spanish) is a critical history of transformational generative models of syntax, from Chomsky’s seminal Logical Structure of Linguistic Theory (1955) to current advances in Minimalism. The primary objective of the book is to explicitly and clearly explain the changes that Generative syntax has undergone in this period, and how these changes relate not only to the internal dynamics of the Generative field (for example, disagreements among different authors as to how to analyse a particular phenomenon), but also the relationship between the transformational generative enterprise and other theories of natural and formal languages (including American and European structuralist formalisms, for natural languages, and Post-Turing formalisms, for formal languages). The book presents the evolution of Generative Grammar as a continuous process, emphasizing the main theoretical developments and proceeding to the specifics while maintaining a sense of historical unity and continuity.

Computer Speech & Language, 2013

LTAG is a rich formalism for performing NLP tasks such as semantic interpretation, parsing, machine translation and information retrieval. Depend on the specific NLP task, different kinds of LTAGs for a language may be developed. Each of these LTAGs is enriched with some specific features such as semantic representation and statistical information that make them suitable to be used in that task. The distribution of these capabilities among the LTAGs makes it difficult to get the benefit from all of them in NLP applications.

My gratitude to the supervisor of this research paper, Pr. Mounia Amrani, is beyond words.

2019

This report presents measurements of the quality of various different dictionaries learned from two different language-learning pipelines; one is the ULL/Kolonin variant, the other is the Linas variant. The ULL/Kolonin dictionaries suggest two major breakthroughs: they suggest how the learning pipeline should be tuned, and they indicate how the pipeline is relatively insensitive to early stages of processing. The Linas variant results indicate that the sheer quantity of training data has a strong impact on the quality of the learned grammars. This report briefly reviews the dictionaries, the algorithms used to obtain them, and measurement results. It is assumed that the reader has a general familiarity with the project. 1 Baseline Measurements This provides the introduction and the baseline to the algorithm, the measurements, and the results.

1998

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact

The Association for Computational Linguistics, 2016

Despite the growing number of Computational Construction Grammar implementations, the field is still lacking evaluation methods to compare grammar fragments across different platforms. Moreover, the hand-crafted nature of most grammars requires profiling tools to understand the complex interactions between constructions of different types. This paper presents a number of evaluation measures, partially based on existing measures in the field of semantic parsing, that are especially relevant for reversible grammar formalisms. The measures are tested on a grammar fragment for European Portuguese clitic placement that is currently under development.