Effective regularisation during training can mean the difference between success and failure for deep neural networks. Recently, dither has been suggested as alternative to dropout for regularisation during batch-averaged stochastic gradient descent (SGD). In this article, we show that these methods fail without batch averaging and we introduce a new, parallel regularisation method that may be ...

This paper describes a modeling and analyzing method for batch sequential control systems by using colored Petri net (CPN) model. For this purpose, we propose systematic procedures to create CPN model from a various information based on ISA S88, i.e., master recipe, equipment information and batch schedule. To examine the effectiveness of the proposed method, we applied it to a two-step polymer...

This paper describes a case of integrating Siebel CRM (Customer Relationship Management) with Finnish Red Cross (FRC) training-site application. The main focus is on evaluating a suggested batch-transfer based integration architecture using use case scenarios and Software Architecture Analysis Method (SAAM) method. The batch approach is compared against synchronous web service based architectur...

—Driven by recent advances in batch Reinforcement Learning (RL), this paper contributes to the application of batch RL to demand response. In contrast to conventional model-based approaches, batch RL techniques do not require a system identification step, which makes them more suitable for a large-scale implementation. This paper extends fitted Q-iteration, a standard batch RL technique, to the...

In this paper a method is proposed called circulant matching method to approximate the superposition of a number of discrete time batch Markovian arrival sources by a circulant batch Markovian process while matching the stationary cumulative distribution and the autocorrelation sequence of the input rate process Special attention is paid to periodic sources The method is applied to the superpos...