The key problem to be faced when building a HMM-based continuous speech recogniser is maintaining the balance between model complexity and available training data. For large vocab...
We propose a new method to group partials produced by each instrument of a polyphonic audio mixture. This method works for pitched and harmonic instruments and is specially adapte...
This note is about work in progress on the topic of "internal type theory" where we investigate the internal formalization of the categorical metatheory of constructive ...
: As supply chains become more dynamic it is important to be able to model them formally as business processes. In particular, there is a need for a sense and respond capability to...
We introduce the model of Markov nets, a probabilistic extension of safe Petri nets under the true-concurrency semantics--this means that traces, not firing sequences, are given a...