The Wikipedia article[1] is actually a pretty decent overview. In short, the simpler versions are extractive -- selecting full sentences from the source, based

The Wikipedia article[1] is actually a pretty decent overview. In short, the simpler versions are extractive -- selecting full sentences from the source, based on supervised or unsupervised machine learning over features such as sentence position, length, similarity to the overall document, etc. [1] http://en.wikipedia.org/wiki/Automatic_summarization http://en.wikipedia.org/wiki/Automatic_summarization