The prevalence of ambiguous acronyms make scientific documents harder to
understand for humans and machines alike, presenting a need for models that can
automatically identify acronyms in text and disambiguate their meaning. We
introduce new methods for acronym identification and disam