============================== Maximum Entropy Discrimination ============================== Marina Meila-Predoviciu We present a general framework for discriminative estimation based on the maximum entropy principle and its extensions. All calculations involve distributions over structures and/or parameters rather than specific settings and reduce to relative entropy protjections. This holds even when the data is not separable within the chosen parametric class, in the context of anomaly detection rather than classication, or when the labels in the training set are uncertain or incomplete. Support vector machines are naturally subsumed under this class and we provide several extensions. We are also able to estimate exactly and efficiently discriminative distributions over tree structures of class-conditional mordels within this framework. Preliminary exprerimental results are indicative of the protential in these techniques. Joint work with Tommi Jaakkola and Tony Jebara.