• Publications
  • Influence
Learning Models for Object Recognition from Natural Language Descriptions
TLDR
This work proposes natural language processing methods for extracting salient visual attributes from natural language descriptions to use as ‘templates’ for the object categories, and applies vision methods to extract corresponding attributes from test images.
Recognising Textual Entailment with Logical Inference
TLDR
This work incorporates model building, a technique borrowed from automated reasoning, and shows that it is a useful robust method to approximate entailment, and uses machine learning to combine these deep semantic analysis techniques with simple shallow word overlap.
The Leeds Arabic Discourse Treebank: Annotating Discourse Connectives for Arabic
We present the first effort towards producing an Arabic Discourse Treebank,a news corpus where all discourse connectives are identified and annotated with the discourse relations they convey as well
Collective Classification for Fine-grained Information Status
TLDR
The task of classifying fine-grained information status and work on written text is introduced and it is claimed that the information status of a mention depends not only on the mention itself but also on other mentions in the vicinity.
Using the Web for Nominal Anaphora Resolution
TLDR
Instead of using handcrafted lexical resources, this work searches the Web with shallow patterns which can be predetermined for the type of anaphoric phenomenon, and achieves state-ofthe-art results.
The Web Library of Babel: evaluating genre collections
TLDR
It is shown that simple character n-grams perform best on current collections because of their ability to generalise both lexical and syntactic phenomena related to genres, and that more research is needed to understand genres on the Web.
Towards a Corpus Annotated for Metonymies: the Case of Location Names
TLDR
A framework for annotating metonymies in domain-independent text that considers the regularity, productivity and underspecification of metonymic usage is described and a fully worked out annotation scheme for location names is presented.
Syntactic Features and Word Similarity for Supervised Metonymy Resolution
TLDR
It is shown that syntactic head-modifier relations are a high precision feature for metonymy recognition but suffer from data sparseness, which is partially overcome by integrating a thesaurus and introducing simpler grammatical features, thereby preserving precision and increasing recall.
Comparing Knowledge Sources for Nominal Anaphora Resolution
TLDR
In studies, the Web-based method alleviated the lexical knowledge gap often encountered in anaphora resolution and handled examples with context-dependent relations between anaphor and antecedent.
Metonymy Resolution as a Classification Task
TLDR
A case study for location names is presented, presenting both a corpus of location names annotated for metonymy as well as experiments with a supervised classification algorithm on this corpus.
...
...