A Summary of Recent Advances in Optical Recognition for Early Music Documents John Ashley Burgoyne Centre for Interdisciplinary Research in Music and Media Technology McGill University Music Technology Area Over the past five years, the Distributed Digital Music Archives and Libraries Laboratory at McGill University has advanced optical recognition technology significantly for early music documents. The project is a good case study for several classical machine learning techniques, including model-discriminant hidden Markov models and maximum à priori adaptation. The project has also established good practises for evaluating preprocessing techniques as part of a larger recognition pipeline and attempted to optimise for client concerns more directly than the traditional trade-off between precision and recall. This talk will summarise these advances with a view toward how they might be applied to music information retrieval more generally.