The process for requirements elicitation has traditionally been based on textual descriptions or graphical models using UML. While these may have worked for the design of desktop-based systems, we argue, that these notations are not adequate for a dialog with mobile end users, in particular for end users in “blue collar” application domains. We propose an alternative modelling technique “Software Cinema” based on the use of digital videos. We discuss one particular example of using Software cinema in the design of a user interface for a navigation system of a mobile end user.
Download (gzipped PDF, 244KB)
In FIPA-style multi-agent systems, agents coordinate their activities by sending messages representing particular communicative acts (or performatives). Agent communication languages must strike a balance between simplicity and expressiveness by defining a limited set of communicative act types that fit the communication needs of a wide set of problems. More complex requirements for particular problems must then be handled by defining domain-specific predicates and actions within ontologies. This paper examines the communication needs of a multi-agent distributed information retrieval system and discusses how well these are met by the FIPA ACL.
Download (gzipped PDF, 322KB)
A modularised connectionist model, based on the Mixture of Experts (ME) algorithm for time series prediction, is introduced. A set of connectionist modules learn to be local experts over some commonly appearing states of a time series. The dynamics for mixing the experts is a Markov process, in which the states of a time series are regarded as states of a HMM. Hence, there is a Markov chain along the time series and each state associates to a local expert. The state transition on the Markov chain is the process of activating a different local expert or activating some of them simultaneously by different probabilities generated from the HMM. The state transition property in the HMM is designed to be time-variant and conditional on the first order dynamics of the time series. A modified BaumÐWelch algorithm is introduced for the training of the time-variant HMM and it has been proved that by EM process the likelihood function will converge to a local minimum. Experiments, with two time series, show this approach achieves significant improvement in the generalisation performance over global models.
Keywords: series prediction, Mixture of Experts, HMM, connectionist model, expectation and maximization, Gauss probability density distribution
Download (gzipped PDF, 450KB)