Neural Networks are primarily concerned with the modelling of biological systems within a mathematical or computer based context. The fundamental component of a network, the neuron, has been the subject of modelling since the middle of the last century. In this paper we explore a new approach to modelling based on integrating the mathematical basis of a model with the rich, interactive processes found in the Empirical philosophy. We introduce a simple notation, N3, for creating models and demonstrate how it can be used to model neural networks through the definition of a simple bottle-neck network. We conclude with a discussion of our approach and ask whether the Empirical approach has benefits when modelling neural networks.
This paper discusses some theories of constructivist learning and the relevance of software models with experiential learning. The applications of Empirical Modelling for two different kinds of teaching models are explored and compared against implementations using procedural languages. Using a series of models of a water supply system, the paper demonstrates the use of an Empirical Modelling approach to teaching abstract concepts and some problems associated with implementing a model of a cyclic system using current empirical modelling tools.
In the field of Empirical Modelling, much has been written and many models created relating to transport systems. However, most of this work has focussed on small-scale systems of a few manually modelled places. There is much to be gained from taking a wider view of a transport system and how it reacts to local conditions. The model accompanying this paper attempts to do so, including various kinds of failures, for a particularly topical example: the London Underground, where there have been an increasing number of security alerts in the wake of the 7th July attacks.
Empirical modelling offers an alternative approach to traditional computing methods. At its centre lies the idea that programs should be built with less emphasis on formal methods, and more around real-world experiences and observations. This means that a distinction must be made between what we believe to be true and what can be observed to be true. Furthermore, the construction of a model should be based on our current understanding of a system and its behaviour, even if that is wrong. With that in mind, this paper will study, with reference to a practical model, whether this approach is in reality a feasible way of learning about and creating a dynamic interactive system.
An investigation into the Empirical Modelling of physical devices, in an educational context, illustrated with the Enigma machine example [Paper , Model ]
This paper investigates the extent to which Empirical Modelling principles are appropriate for modelling physical devices, and the advantages and disadvantages of doing so, illustrated with a `prototype' model of the series of German encryption machines known as Enigma. The paper discusses how effectively such models of physical devices can be used in an educational environment, any considerations that may need to be taken into account when attempting such a task, and ultimately concludes that despite a few weaknesses, the potential for creating a useful model that aids learning is high.
This paper examines the benefits of using an Empirical Modelling approach to simulating the movement of a collection of people out of an area during an emergency, using an original model developed with the tkeden tool. It is found that such a model can benefit from the use of EM, since the need to allow the user to experiment is met by a coherent, open environment which can cope well with change because of the use of dependency. Shortcomings are found in EM's dependency notation, and language extensions are given to address these issues.
Grids are broadcast as the framework to perform rapid data analysis. This is often required to gain a competitive advantage in industry. The speed of any such network, however, depends largely on its architecture. In this paper, I describe a model that allows users to explore the effects on a Grid's perfor ance of altering its arrangement. I start by discussing the relevant subject area, followed by a critique of empirical principles and the tkeden toolkit n modelling this state.
This paper argues for an agent orientated approach to modelling human computer interaction, through the use of widgets. These are small fragments of a script that are used to help a human interact with a computer model. An example implementation, the Agent Orientated Widget Toolkit, is presented along with its notation. This is then compared to an object orientated approach to writing a 'user Interface' toolkit.
Empirical Modelling for Education: Modelling the Routing Information Protocol (RIP-1) [Paper , Model ]
Teaching of the principles and operation of complex systems (particularly in science-based subjects) often demands that students be presented with a simplified model of a system, with irrelevant details omitted so that only those relevant to a particular educational objective remain. This paper suggests problems that may be encountered by teachers in such cases, and proposes that when combined with traditional teaching methods, Empirical Modelling provides an effective solution that can be utilised by both teachers and students. Teaching of the Routing Information Protocol (RIP-1) is used as an example to demonstrate the strengths and limitations of creating educational models with the EDEN software tool.
Although Newtonian Mechanics is known to not be an entirely accurate model describing the motion of bodies, it is still widely used due to its relative simplicity and adequate level of accuracy for many applications. This report discusses the issues surrounding the development of a model as an educational aid for learning Newtonian Mechanics, including the characteristics a successful model would possess.
Charles Babbage's Difference Engine is commonly considered to be the world's first computer. As such it is an important part of any Computing teaching. Though much literature exists on Babbage, the construction of the engine and the algorithm it computes, little has been written about the actual operation of the machine. Constructionism is well known in the field of teaching and learning. This paper will attempt to identify the relative merits of constructionism in creating a model of the Difference Engine for educational purposes, describe the model and process and discuss this approach with regards to its relationship to Empirical Modelling.
This paper discusses the need and suitability for an Empirical Modelling approach for the provision of business optimisation support. This is with particular reference to the fast food industry, with a model of this environment being created. Its limitations, and those of EM as a whole in this area are also discussed.
Su Doku, "The Rubik's cube of the 21st century" is the faster growing puzzle phenomenon in the world, replacing traditional crosswords, alphawords and lateral thinking games in nearly all major newspapers, Smith (2005). A craze that is almost impossible to ignore, with clubs, worldwide tournaments, books, software and even television programmes dedicated to the puzzle. In this paper we introduce the concept of the Doku-games, a family of puzzles that share many properties with Su Doku, and outline a script of definitions to create a comprehensive Doku-game model.
An Empirical Modelling Approach to Educational Technology in Learning about Business [Paper , Model ]
Continuous change in technology and the awareness of a new approach for education led to a `revolution in education' in the past. Educational Technology (ET) is the proposed solutions for these problems as a result of rethinking education. One implementation of ET is using Empirical Modelling (EM) techniques. Learning through interaction with the computer plays an important role in thinking about education. Therefore a business model is developed to illustrate the learning process.
This paper explores some of the dependencies involved in the operations of a typical hospital's Accident and Emergency ward, and how agents can help define the behaviour of processes involved in its operations. An attempt is made at agentification using the LSD notation. A model of the A&E ward built using tkeden is described and compared against a similar model built using SIMUL8, a popular modelling software.
The paper details an approach to Natural Language Processing (NLP) based on the principles of Empirical Modelling (EM). The suitability of EM as a tool for working with natural languages is discussed as well as how NLP can integrate EM ideals. A definitive notation for making use of natural language in EM IS presented, followed by an overview of an implementation, paying particular regard to how it makes use of the features of EM.