This paper presents a novel learning algorithm for structured classification, where the task is to predict multiple and interacting labels (multilabel) for an input object. The problem of finding a large-margin separation between correct multilabels and incorrect ones is formulated as a linear program. Instead of explicitly writing out the entire problem with an exponentially large constraint s...

The author presents solutions to input-to-state stabilization and integral input-to-state stabilization problems for nonlinear systems based on the novel concept of state-dependent scaling design. Both state-feedback and output-feedback controllers are constructed in a uni ed way. The method provides global solutions whenever the system is in the strict-feedback or output-feedback form. The pap...

The development of a spoken dialogue system is a complex process involving the integration of several component technologies. Various toolkits and authoring environments have been produced that provide assistance with this process. This paper reports on several projects involving CSLU’s RAD (Rapid Application Developer) and critically evaluates the applicability of state transition diagrams for...

Many tasks involve predicting a large number of variables that depend on each other as well as on other observed variables. Structured prediction methods are essentially a combination of classification and graphical modeling. They combine the ability of graphical models to compactly model multivariate data with the ability of classification methods to perform prediction using large sets of inpu...

In structured prediction, it is standard procedure to discriminatively train a single model that is then used to make a single prediction for each input. This practice is simple but risky in many ways. For instance, models are often designed with tractability rather than faithfulness in mind. To hedge against such model misspecification, it may be useful to train multiple models that all are a ...

Many language generation tasks require the production of text conditioned on both structured and unstructured inputs. We present a novel neural network architecture which generates an output sequence conditioned on an arbitrary number of input functions. Crucially, our approach allows both the choice of conditioning context and the granularity of generation, for example characters or tokens, to...

State-of-the-art TCAD software like the multi-dimensional device and circuit simulator MINIMOSNT require a huge amount of information in addition to the device input data to control several different complex modules and tasks. This information is normally hierarchically structured containing arbitrary cross-relations and dependencies that include, e.g., material properties, circuit descriptions...