of the work presented in this thesis has been previously published as listed below. Although some of these papers have co-authors, the work appearing in this thesis is entirely my own, with the exception of parts of chapter 3, which presents work jointly carried out by myself and Adrian Thompson. The respective contributions to this work will be explicitly stated at the beginning of the chapter. List of Previous Publications Kuntz, P., Layzell, P., & Snyers, D. (1997). A Colony of Ant-like Agents for Partitioning

The CAM-Brain Machine (CBM) is a hardware implementation of a brain-inspired, recurrent, digital neural network. It is an experimental machine composed of reconfigurable (evolvable) hardware, capable of training and evaluating cellular automata based neural network modules directly in silicon.

. This article introduces ATR's "CAM-Brain Machine" (CBM), an FPGA based piece of hardware which implements a genetic algorithm (GA) to evolve a cellular automata (CA) based neural network circuit module (of approximately 1000 neurons) in about a second (i.e. a complete run of a GA, with 10,000s of circuit growths and performance evaluations) . Up to 32000 of these modules (each of which is evolved with a humanly specified function) can be downloaded into a large RAM space, and interconnected according to humanly specified artificial brain architectures. This RAM, containing an artificial brain with up to 40 million neurons, is then updated by the CBM at a rate of 150 Billion CA cells per second. Such speeds should enable real time control of robots and hopefully the birth of a new research field that we call "brain building". The first such artificial brain (to be built by ATR in 1999) will be used to control the behaviors of a life sized robot kitten called "Robokoneko". 1 Introducti...

This paper presents an algorithm which converts an arbitrary analog time-varying signal into a digital spiketrain (a bit string of 0's interspersed with 1's), where the information is contained in the spacing between the spikes. This conversion is an important ingredient in the CAM-Brain Project, as it allows the user of ATR's CAM-Brain Machine (CBM) to think entirely in terms of analog signals, and not the more abstract, visually rather meaningless, spiketrains. The SPIKER conversion completes a package of which makes the evolution of individual CBM based neural network modules easier to think about and to accomplish. Keywords: evolutionary neural networks, hardware implementation, CAM-Brain project, CoDi neural model, information representation schemes 1 Introduction ATR's CAM-Brain Project [1] aims to build a largescale brain-like neural network system in hardware. The essential ingredient in this project is a special piece of hardware, based on Xilinx XC6264 FPGAs which grow and ...

This paper introduces conceptual problems Abstract - This paper introduces conceptual problems that will arise in the next 10-20 years as electronic that will arise in the next 10-20 years as electronic circuits reach nanometer scale, i.e. the size of molecules. circuits reach nanometer scale, i.e. the size of molecules. Such circuits will be impossible to make perfectly, due to Such circuits will be impossible to make perfectly, due to the inevitable fabrication faults in chips with an the inevitable fabrication faults in chips with an Avogadro number of components. Hence they will need Avogadro number of components. Hence they will need to be constructed so that they are robust to faults. They to be constructed so that they are robust to faults. They will also need to be (as far as possible) reversible circuits, will also need to be (as far as possible) reversible circuits, to avoid the heat dissipation problem if bits of to avoid the heat dissipation problem if bits of information are routinely wiped out during the information are routinely wiped out during the computational process. They will also have to be local if computational process. They will also have to be local if the switching times reach femto-seconds, which is possible the switching times reach femto-seconds, which is possible now with quantum optics. This paper discusses some of now with quantum optics. This paper discusses some of the conceptual issues involved in trying to build circuits the conceptual issues involved in trying to build circuits that satisfy all these criteria, i.e. that they are that satisfy all these criteria, i.e. that they are robust, robust, reversible and loca reversible and local. We propose an evolutionary l. We propose an evolutionary engineering based model that meets all t...

This paper presents a Learning Classifier System (LCS) where each classifier condition is represented by a spiking neural network. Adaptive behavior is realized through the use of self-adaptive parameters and neural constructivism, providing the system with a flexible knowledge representation. The approach allows for the evolution of networks of appropriate complexity to solve a continuous maze environment, here using discrete-valued actions. It is shown that the neural LCS is capable of developing optimal solutions to the Agent navigation tasks serve as a well-established test bed for learning systems. Typical tasks involve an agent, which is initially situated randomly within a maze environment, using sensory readings to navigate to a goal state; arrival at the goal state triggers a reward. These kinds of navigation tasks can be broadly defined as either discrete (e.g., [26]) or continuous (e.g., [4]).