Question Answering Based on Semantic Graphs
author: Delia Rusu, Artificial Intelligence Laboratory, Jožef Stefan Institute
author: Blaž Fortuna, Artificial Intelligence Laboratory, Jožef Stefan Institute
author: Dunja Mladenić, Artificial Intelligence Laboratory, Jožef Stefan Institute
author: Marko Grobelnik, Artificial Intelligence Laboratory, Jožef Stefan Institute
published: May 27, 2009, recorded: April 2009, views: 5980
Report a problem or upload filesIf you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.
Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.
In this paper we present a question answering system supported by semantic graphs. Aside from providing answers to natural language questions, the system offers explanations for these answers via a visual representation of documents, their associated list of facts described by subject – verb – object triplets, and their summaries. The triplets, automatically extracted from the Penn Treebank parse tree obtained for each sentence in the document collection, can be searched, and we have implemented a question answering system to serve as a natural language interface to this search. The vocabulary of questions is general because it is not limited to a specific domain, however the questions's grammatical structure is restricted to a predetermined template because our system can understand only a limited number of question types. The answers are retrieved from the set of facts, and they are supported by sentences and their corresponding document. The document overview, comprising the semantic representation of the document generated in the form of a semantic graph, the list of facts it contains and its automatically derived summary, offers an explanation to each answer. The extracted triplets are further refined by assigning the corresponding co referenced named entity, by resolving pronominal anaphors, as well as attaching the associated WordNet synset. The semantic graph belonging to the document is developed based on the enhanced triplets while the document summary is automatically generated from the semantic description of the document and the extracted facts.
Link this pageWould you like to put a link to this lecture on your homepage?
Go ahead! Copy the HTML snippet !