This work targeted at abstract references to events. Various sources of knowledge (such as position, lexical, dictionary, syntax and semantic knowledge) were applied.
This work started from a prelimanry exploration on event pronoun resolution which was published on CoLing 2010 at Beijing, China.
The second part extended to event noun phrases resolution which to appear in EMNLP 2010 at MIT, Cambridge, MA, USA.
This work targeted at resolving anaphors pre-modified with keyword "other" and "another". Traditional rule-based linguistics features were combined with automatically mined patterns from the web.
This work was published on CoLing 2008 at Manchester, UK.
Natural Language Processing(NLP) is is a field of computer science and linguistics concerned with the interactions between computers and human (natural) languages.
In linguistics, anaphora is an instance of an expression referring to another. Anaphora resolution is the task of finding out what an anaphor is referring to. In linguistics, co-reference occurs when multiple expressions in a sentence or document have the same referent. Co-reference resolution is the task to idetify co-referred expressions. The difference between the two terms are 1. anaphor requires the antecedent information for a correct interpretation while co-reference is not necessary so; 2. co-reference enforces identy relation among co-referred expressions while anaphora can have various types of relations.
Bridging Anaphora is a general term for the non-identical relations in anaphora phenomenon. Some instances for bridging include part-whole relation (e.g. car-wheel), class-instance relation(e.g. fruit-apple) and attribute relation (e.g. house-size).
Event Anaphora/Co-reference is an anaphoric/co-referring instance which the antecedent is an event. In conventional anaphora phenomenon, most of anaphors are referring to real world objects. Event anaphora is referring to an abstract entity which is an event.
Some useful tools, parsers and corpus I used in my study.
Machine Learning Tools
Parsers
Corpus