Natural language processing tools assisted by human experts show promise in interpretation of electronic medical records (EMRs). While manual review is generally used to confirm or reject NLP results, there is a scope for using such feedback to improve upon the machine-learned models. In this project, I explore the design space to facilitate this process and suggest novel interfaces that can be used by the experts (medical practitioners) for providing feedback. This would eventually be a part of a larger system that learns by combining NLP predictions with human knowledge. I will present a review of the recent work in interactive machine learning and how we can adapt some of these techniques for our tool.