Recent work in spoken language modeling shows the possibility of learning a unsupervisedly from raw audio without any text labels. The approach relies first on transforming into sequence discrete units (or pseudo-text) and then training model directly such pseudo-text. Is bottleneck necessary, potentially introducing irreversible errors encoding speech signal, or could we learn at all? In this ...

Stand-alone modeling and distributed modeling show considerable differences in methodology and amount of modeling effort needed. An approach is discussed to design and develop distributed models with the ease of stand-alone modeling and without the overhead usually associated with distributed modeling. The main objective is to obtain interchangeability of a stand-alone model and its distributed...

Acknowledgements: Justin Tobias provided many useful references and comments on the third section of this paper. provided many useful comments and references. Of course, none of these people are responsible for any of the errors or omissions in this paper.

by Magne Myrtveit Normally, we try to aggregate discreteness in System Dynamics (SD) models by making continuous flows. This is a good modeling practice, but in some cases discreteness may have a significant influence on system behavior (see note). Consequently, we need to model such discreteness to explore its impact on behavior. The System Dynamics (SD) technology was originally designed with...