can a computer "read" an online blog and understand it? several concordia computer scientists are
1 to get closer to that goal. leila kosseim, associate professor in concordia's
2 of engineering and computer science, and a recently-graduated doctoral student, shamima mithun, have developed a system called blogsum that has potentially vast applications. it allows an organization to pose a question and then find out how a large number of people talking online would respond. the system is capable of
3 things like consumer preferences and voter intentions by sorting through websites, examining real-life self-expression and conversation, and producing summaries that focus exclusively on the original question.
"huge quantities of electronic texts have become easily available on the internet, but people can be overwhelmed, and they need help to find the real content hiding in the mass of information," explains kosseim, one of the lead researchers at concordia's computational
4 laboratory (clac lab).
5 informally-written language poses unique challenges compared to analyzing, for example, a news article. blogs,
6 and the like contain opinions, emotions and
7, not to mention spelling errors and poor grammar. a summarization tool must address two particular problems, question
8 (sentences that are not relevant to the main question), and
9 incoherence, (sentences in which the intent of the writer is unclear).
blogsum met these challenges with demonstrable efficiency. the researchers developed and tested their tool by examining a set of blogs and review sites. blogsum used "discourse relations" to
10 the data -- ways of filtering and ordering sentences into coherent summaries. blogsum was measured against prior computational rankings and achieved mostly superior results. in addition, it was evaluated by actual human subjects, who also found it to be superior. summaries produced by blogsum reduced question irrelevance and discourse incoherence, successfully
11(蒸馏的) large amounts of text into highly readable summaries.
this study is an example of natural language processing (nlp), in which concordia, through the clac lab, is a leader. nlp stands at the
12 of artificial intelligence and linguistics, seeking to enable computers to
13 meaning from human language.
"the field of natural language processing is starting to become fundamental to computer science, with many everyday applications -- making search engines find more relevant documents or making smart phones even smarter," explained kosseim.