We study weighted programming, a programming paradigm for specifying mathematical
models. More specifically, the weighted programs we investigate are like usual imperative …

We present algorithms, implemented as an extension to the OpenFst library, that yield a
class of transducers that encode linear models for structured inference tasks like …

Huge neural autoregressive sequence models have achieved impressive performance
across different applications, such as NLP, reinforcement learning, and bioinformatics …

We describe a simple and efficient algorithm to disambiguate non-functional weighted finite
state transducers (WFSTs), ie to generate a new WFST that contains a unique, best-scoring …

Central to natural language understanding and generation is the ability to accurately model
the semantics of text. We can frame understanding as parsing a semantic form from a given …

Die Arbeit folgt der These, dass der Bereich der Constraint-Verarbeitung und der Bereich
der endlichen Techniken ein großes Potenzial haben, sich gegenseitig mit neuen …

We introduce a lattice determinization algorithm that can operate incrementally. That is, a
word-level lattice can be generated for a partial utterance and then, once we have …

Recent advances in speech recognition technology and the growing societal adoption of
mobile devices have led to an unprecedented demand for natural spoken interfaces in a …