... conference on Artificial intelligence (vol 1), AAAI ’94, pages 779–784, Menlo Park, CA, USA American Association for Artificial Intelligence Lavie, A and Agarwal, A (2007) Meteor: an automatic metric ... English speakers for grammar correction Using only these data sets, we can train our noisy channel model, as we have shown using a bigram language model, and a wFST for our noise model We have also ... ε:ε/1 an:an/0.4 4.3 Learning noise model parameters cat:cat/1 ear:ear/1 ear:ear/1 cat:cat/1 an:an/1 a: an/0.05[0,0.05] a: a/0.95[0.95,0] To achieve maximum performance, we wish to learn the parameters...