Teijiro Isokawa

Fault-tolerance in nanocomputers: a cellular array approach

Asynchronous cellular arrays have gained attention as promising architectures for nanocomputers, because of their lack of a clock, which facilitates low power designs, and their regular structure, which potentially allows manufacturing techniques based on molecular self-organization. With the increase in integration density comes a decrease in the reliability of the components from which computers are built, and implementations based on cellular arrays are no exception to this. This paper advances asynchronous cellular arrays that are tolerant to transient errors in up to one third of the information stored by its cells. The cellular arrays require six rules to describe the interactions between the cells, implying less complexity of the cells as compared to a previously proposed (nonfault-tolerant) asynchronous cellular array that employs nine rules.

ASSOCIATIVE MEMORY IN QUATERNIONIC HOPFIELD NEURAL NETWORK

Associative memory networks based on quaternionic Hopfield neural network are investigated in this paper. These networks are composed of quaternionic neurons, and input, output, threshold, and connection weights are represented in quaternions, which is a class of hypercomplex number systems. The energy function of the network and the Hebbian rule for embedding patterns are introduced. The stable states and their basins are explored for the networks with three neurons and four neurons. It is clarified that there exist at most 16 stable states, called multiplet components, as the degenerated stored patterns, and each of these states has its basin in the quaternionic networks.

Quaternion Neural Network and Its Application

Quaternion neural networks are models of which computations in the neurons is based on quaternions, the four-dimensional equivalents of imaginary numbers. This paper shows by experiments that the quaternion-version of the Back Propagation (BP) algorithm achieves correct geometrical transformations in color space for an image compression problem, whereas real-valued BP algorithms fail.

Fundamental Properties of Quaternionic Hopfield Neural Network

Associative memory by Hopfleld-type recurrent neural networks with quaternionic algebra, called quaternionic Hopfield neural network, is proposed in this paper. The variables in the network are represented by quaternions of four dimensional hypercomplex numbers. The neuron model, the energy function, and the Hebbian rule for embedding patterns into the network are introduced. The properties of this network are analyzed concretely through examples of the network with 3 and 4 quaternion neurons. It is demonstrated that there exist fixed attractors in the network, i.e., the pattern association from test pattern close to a stored pattern is possible in the quaternionic network, as in real-valued Hopfleld networks.

Self-Organizing Map Based Data Detection of Hematopoietic Tumors

Data detection based on self organizing maps is presented for hematopoietic tumor patients. Learning data for the maps are generated from the screening data of examinees. The incomplete screening data without some item values is then supplemented by substituting averaged non-missing item values. In addition, redundant items, which are common to all the data and tend to have an unfavorable influence on data detection, are eliminated by a genetic algorithm and/or an immune algorithm. It is basically judged, by observing the label of a winner neuron in the map, whether the data presented to the map belongs to the class of hematopoietic tumors. Some experimental results are provided to show that the proposed methods achieve the high probability of correctly identifying examinees as hematopoietic tumor patients.

Commutative quaternion and multistate Hopfield neural networks

This paper explores two types of multistate Hopfield neural networks, based on commutative quaternions that are similar to Hamiltons quaternions but with commutative multiplication. In one type of the networks, the state of a neuron is represented by two kinds of phases and one real number. The other type of the networks adopts the decomposed form of commutative quaternion, i.e., the state of a neuron consists of a combination of two complex values. We have investigated the stabilities of these networks, i.e., the energies monotonically decreases with respect to the changes of the network states.

Reconstruction and simulation for three-dimensional morphological structure of insect neurons

We present a system for the reconstruction three-dimensional morphological structure of a neuron from a sequence of tomographic images acquired by a confocal laser scanning microscope. In this system, the branching structure and diameter of dendrites are extracted by the Single-Seed Distance Transform method. Compartmental neuron models are reconstructed using the morphological structure detected by our system. In order to analyze electrical properties, a model description for the neuronal simulator, NEURON, is generated automatically. The effectiveness of the proposed system is shown by application to the reconstruction of interneurons in an antennal lobe of silkworm moths.

Defect-tolerance in cellular nanocomputers

For the manufacturing of computers built by nanotechnology, defects are expected to be a major problem. This paper explores this issue for nanocomputers based on cellular automata. Known for their regular structure, such architectures promise cost-effective manufacturing based on molecular self-organization. We show how a cellular automaton can detect defects in a self-contained way, and how it configures circuits on its cells while avoiding the defects. The employed cellular automaton is asynchronous, i.e., it does not require a central clock to synchronize the updates of its cells. This mode of timing is especially suitable for the high integration densities of nanotechnology implementations, since it potentially causes less heat dissipation.

Emergence of flocking behavior based on reinforcement learning

Grouping motion, such as bird flocking, land animal herding, and fish schooling, is well-known in nature. Many observations have shown that there are no leading agents to control the behavior of the group. Several models have been proposed for describing the flocking behavior, which we regard as a distinctive example of the aggregate motions. In these models, some fixed rule is given to each of the individuals a priori for their interactions in reductive and rigid manner. Instead of this, we have proposed a new framework for self-organized flocking of agents by reinforcement learning. It will become important to introduce a learning scheme for making collective behavior in artificial autonomous distributed systems. In this paper, anti-predator behaviors of agents are examined by our scheme through computer simulations. We demonstrate the feature of behavior under two learning modes against agents of the same kind and predators.

Quaternionic multistate Hopfield neural network with extended projection rule

The aim of this paper is to investigate storing and recalling performances of embedded patterns on associative memory. The associative memory is composed of quaternionic multistate Hopfield neural network. The state of a neuron in the network is described by three kinds of discretized phase with fixed amplitude. These phases are set to discrete values with arbitrary divide size. Hebbian rule and projection rule are used for storing patterns to the network. Recalling performance is evaluated through storing random patterns with changing the divide size of the phases in a neuron. Color images are also embedded and their noise tolerance is explored.