Data Set Used
Declaration I declare that the work described in this thesis has not been submitted for a degree at any other university, and that the work is entirely my o wn. I agree that the library in Trinity College Dublin may lend or copy this thesis upon request. Acknowledgements This thesis owes so much t o t h e great people who helped me learn and who made the… (More)
This paper is concerned with some aspects of the semantics of modal verbs in English, in particular the phenomenon of modal subordination and the constraints on interpretation of successive modal sentences in a discourse. I propose a solution in the framework of dynamic semantics. The information state is taken to be a set of possibilities which are… (More)
Many plain text information hiding techniques demand deep semantic processing, and so suffer in reliability. In contrast, syntactic processing is a more mature and reliable technology. Assuming a perfect parser, this paper evaluates a set of automated and reversible syntactic transforms that can hide information in plain text without changing the meaning or… (More)
We introduce the formal underpinnings of our theory of non-classical feature structures. The resulting expanded universe of feature structures has direct impfications for robust parsing for linguistic theories founded upon feature theory. We present an implementation of a robust chart parser for Head-driven Phrase Structure Grammar (HPSG). The problem of… (More)
Head-Driven Phrase Structure Grammar (HPSG), a unification-based formal language for describing linguistic phenomena, has a declarative semantics which makes it amenable to specification as a logic program. The HPSG formalism has undergone significant modification, becoming more declarative and incorporating greater lexicalization, since Proudian and… (More)
Error-Correcting Output Coding (ECOC) is a general framework for multiclass text classification with a set of binary classifiers. It can not only help a binary classifier solve multi-class classification problems, but also boost the performance of a multi-class classifier. When building each individual binary classifier in ECOC, multiple classes are… (More)
We present three natural language marking strategies based on fast and reliable shallow parsing techniques, and on widely available lexical resources: lexical substitution, adjective conjunction swaps, and relativiser switching. We test these techniques on a random sample of the British National Corpus. Individual candidate marks are checked for goodness of… (More)
In this paper we present the approach we took in our participation to the PAN 2013 Author Profiling task. It is an adaptation of our system submitted for author identification, assuming that a profile category (authors belonging to the same gender and age group categories) can be analyzed in the same way as an author's style.
In this paper we present the system we submitted to the PAN 2015 competition for the author verification task. We consider the task as a supervised classification problem, where each case in a dataset is an instance. Our approach combines the output from multiple learners using basic stacked generalization. The individual learners are obtained using five… (More)
In this paper we present the approach we took in our participation to the PAN 2013 Author Identification task. It relies on a complex process to select the features which represent the author's writing, using potentially multiple statistics and distance measures computed from the training set.