Infomagnetism and Sentence Generation

L attraction was introduced by Doug Beeferman, Adam Berger and John Lafferty in a 1997 conference paper [1], and was picked up by Deniz Yuret in his 1998 thesis which focused on using lexical attraction to discover a link-grammar for English utterances [2]. Searching for "lexical attraction" on Google reveals links to these papers, and more. Lexical attraction is, however, nothing more than the pointwise mutual information between two quantifiable entities [3]. The concept of mutual information has long been used as a measure of the correlation between the two random variables due to the fact that it reflects the reduction of uncertainty of one random variable due to the knowledge of the other.