Tatsumi Furuya

Hardware Evolution at Function Level

This paper describes a function-level Evolvable Hardware (EHW). EHW is hardware which is built on programmable logic devices (e.g. PLD and FPGA) and whose architecture can be reconfigured by using a genetic learning to adapt to new unknown environments in real time. It is demonstrated that the function-level hardware evolution can attain much higher performances than the gate-level evolution, in neural network applications (e.g. two-spiral). VLSI architecture of the functionbased FPGA dedicated to function level evolution is also described.

Evolvable Hardware and Its Applications to Pattern Recognition and Fault-Tolerant Systems

This paper describes Evolvable Hardware (EHW) and its applications to pattern recognition and fault-torelant systems. EHW can change its own hardware structure to adapt to the environment whenever environmental changes (including hardware malfunction) occur. EHW is implemented on a PLD(Programmable Logic Device)-like device whose architecture can be altered by re-programming the architecture bits. Through genetic algorithms, EHW finds the architecture bits which adapt best to the environment, and changes its hardware structure accordingly.

Evolvable hardware with genetic learning

This paper describes Evolvable Hardware (EHW) with genetic learning. EHW is hardware which is built on programmable logic devices (e.g. PLD and FPGA) and whose architecture can be reconfigured by using genetic learning to adapt to the new environment. There are two types of hardware evolutions; gate-level and function-level. As examples of gate-level evolution, a pattern recognition system and a welding robot controller are described. Then, function-level EHW is introduced. It is demonstrated that function-level hardware evolutions can attain high performances as in neural network applications (e.g. two spirals). New FPGA architecture for function-level evolution is also described.

IXM2: a parallel associative processor

This paper describes a parallel associative processor, lXM2, developed mainly for semantic network processing. IXM2 consists of 64 associative processors and 9 network processors, having a total of 256K words of associative memory. The large associative memory enables 65,536 semantic network nodes to be processed in parallel and reduces the order of algorithmic complexity to O( 1) in,basic semantic net operations. It is shown that IXM2 has computing power comparable to that of a Connection Machine. Programming for lXM2 is performed with the knowledge representation language IXL, a superset of Prolog, so that IXM2 can be utilized as a back-end to AI workstations.

Evolutionary programming for mix design

Abstract An evolution method for the solution of a mix design problem is proposed. It is an application of genetic algorithms in which the ratio of ingredients have evolved. In this paper, the evolution algorithm, new crossover methods and mutation schemes are introduced. The usual crossover operation was not used, instead one field of a superior chromosome was copied to an inferior chromosome. A mutation was generated by a combination of uniform random number and normal distribution random number and the mutation rate was very high ( 1 2 ). This approach can be applied to non linear optimization problems. The method has been successfully applied to the actual mix problem.

An Extended Method of Higher-order Local Autocorrelation Feature Extraction for Classification of Histopathological Images

In histopathological diagnosis, a clinical pathologist discriminates between normal tissues and cancerous tissues. However, recently, the shortage of clinical pathologists is posing increasing burdens on meeting the demands for such diagnoses, and this is becoming a serious social problem. Currently, it is necessary to develop new medical technologies to help reduce their burdens. Therefore, as a diagnostic support technology, this paper describes an extended method of HLAC feature extraction for classification of histopathological images into normal and anomaly. The proposed method can automatically classify cancerous images as anomaly by using an extended geometric invariant HLAC features with rotation- and reflection-invariant properties from three-level histopathological images, which are segmented into nucleus, cytoplasm and background. In conducted experiments, we demonstrate a reduction in the rate of not only false-negative errors but also of false-positive errors, where a normal image is falsely classified as an image with an anomaly that is suspected as being cancerous.

Proposal for LDPC Code Design System Using Multi-Objective Optimization and FPGA-Based Emulation

The paper proposes a low density parity check (LDPC) code design system to facilitate the design of communication systems using LDPC codes for error correction. The proposed LDPC code design system has three advantages (utilization of MOGA to search codes, speed enhancement achieved through parallelization and FPGAs, and employment of more precise simulation models) and solves problems encountered when LDPC codes are used in practical applications. Preliminary evaluation results for the proposed system are presented, which demonstrate that the system can function successfully.

Coevolution in recurrent neural networks using genetic algorithms

This paper describes an investigation into the effectiveness of a lookahead model based on recurrent neural networks. An action network and an internal model of the environment are incorporated into the recurrent neural network; lookahead planning is performed while configuring the action network through learning in the internal model. A genetic algorithm is applied to the design of the neural networks. The effectiveness of this model is evaluated by applying it to the game of “tic-tac-toe,” and the following result is obtained. It is possibly more effective to perform learning in the internal model by learning algorithms than by memorizing input-output correspondences.

Cost coefficient control method for solving optimization problems on Hopfield-type neural networks

When solving optimization problems on a Hopfield-type neural network, a constraint coefficient and cost coefficient of an energy function should be determined appropriately. Until recently, the values of these coefficients were decided based on experience and trial and error. Therefore, solutions that satisfy the constraints could not be obtained and the quality of the solutions was not good. In order to avoid this problem, we propose a method to control cost coefficient values automatically while keeping a constraint coefficient to be constant. We applied this method to the Travelling Salesman Problem, and obtained near-optimal solutions more efficiently than other approaches. The proposed algorithm is very effective especially for the difficult city allocations.

Evolvable Optical Systems and Their Applications

This paper describes evolvable optical systems and their applications developed at the National Institute of Advanced Industrial Science and Technology (AIST) in Japan. Three evolvable optical systems are described: (1) an evolvable fiber alignment system, (2) an evolvable interferometer system, and (3) an evolvable femtosecond laser system. As the micron-meter resolution alignment of optical components usually takes a long time, to overcome this time problem, we propose 3 systems that can automatically align the positioning of optical components by genetic algorithms in very short times compared to conventional systems. In the evolvable fiber alignment system, the positioning of a fiber can be aligned automatically according to 5 degrees of freedom (DOF) in three minutes. In the evolvable interferometer system, the optimal positioning of the plane mirrors is determined automatically. And, in the evolvable femto-second laser system, the positioning of laser components can be aligned automatically within thirty minutes; something which in conventional system would take technicians more than five days to achieve. Moreover, the automatic alignment system makes possible the compact implementation of optical systems.