Papers - YAMAMORI Kunihito
-
Classification of species by information entropy and visualization by self-organizing map Reviewed
Kentaro NISHIMUTA, Ikuo YOSHIHARA, Kunihito YAMAMORI, Moritoshi YASUNAGA
Proc. Sixteenth International Symposium on Artificial Life and Robotics 371 - 374 2011.1
Language：English Publishing type：Research paper (international conference proceedings)
-
Asynchronous migration for parallel genetic programming on computer cluster with multi-core processers Reviewed
Shingo KUROSE, Kunihito YAMAMORI, Masaru AIKAWA, Ikuo YOSHIHARA
Proc. Sixteenth International Symposium on Artificial Life and Robotics 367 - 370 2011.1
Language：English Publishing type：Research paper (international conference proceedings)
-
Neural network with exponential output neuron for estimation of physiological activities from protein expression levels Reviewed
Kazuhiro KONDO, Kunihito YAMAMORI, Ikuo YOSHIHARA
Proc. Sixteenth International Symposium on Artificial Life and Robotics 363 - 366 2011.1
Language：English Publishing type：Research paper (international conference proceedings)
-
A design of Self-defect-compensatable hardware neuron for multi-layer neural networks
Yamamori K., Tashiro K., Kusano M., Yoshihara I.
Proceedings - IEEE International Symposium on Defect and Fault Tolerance in VLSI Systems 82 - 89 2010.12
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：Proceedings - IEEE International Symposium on Defect and Fault Tolerance in VLSI Systems
Neural network has a problem that learning time becomes so long for real world problems. To achieve fast learning, some researchers proposed to implement a neural network into Wafer Scale Integration (WSI). Since WSI uses one wafer as a parallel computer, a part of defect leads entire system fault. Therefore a defect compensation method is necessary to implement a neural network into WSI. Partial Retraining (PR) method has proposed as one of the defect compensation methods for neural network. However PR method is not verified whether it will perform well on digital hardware or not. It is also not clear how much is circuit required. In this paper we report a design of self-defect-compensatable neuron with PR method by VHDL, and evaluate it by simulations. © 2010 IEEE.
DOI： 10.1109/DFT.2010.17
-
Kuroda M., Yamamori K., Munetomo M., Yasunaga M., Yoshihara I.
Artificial Life and Robotics 15 ( 4 ) 547 - 550 2010.12
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：Artificial Life and Robotics
This article proposes a novel crossover operator of hybrid genetic algorithms (HGAs) with a Lin-Kernighan (LK) heuristic for solving large-scale traveling salesman problems (TSPs). The proposed crossover, tentatively named sub-tour recombination crossover (SRX), collects many short sub-tours from both parents under some set of rules, and reconnects them to construct a new tour of the TSP. The method is evaluated from the viewpoint of tour quality and CPU time for ten well-known benchmarks, e. g., dj38, qa194, ..., ch71009. tsp, in the TSP website of the Georgia Institute of Technology. We compare the SRX with three conventional crossover operators, a variant of the maximal preservative crossover operator (MPX3), a variant of the greedy sub-tour crossover operator (GSX2), and a variant of the edge recombination crossover operator (ERX6), and show that the SRX succeeded in finding a better solution and running faster than the conventional methods mentioned above. © 2010 International Symposium on Artificial Life and Robotics (ISAROB).
-
Kuroda M., Yamamori K., Munetomo M., Yasunaga M., Yoshihara I.
Proceedings of the 15th International Symposium on Artificial Life and Robotics, AROB 15th'10 828 - 831 2010.12
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：Proceedings of the 15th International Symposium on Artificial Life and Robotics, AROB 15th'10
In this paper, we propose a novel crossover operator for solving the traveling salesman problem (TSP) with a Hybrid Genetic Algorithm (HGA) involving Lin-Kernighan (LK) heuristic for local search. We call the crossover operator Sub-tour Recombination Crossover (SRX) which divides each tour of the parents into many sub-tours under some rules and reconnects sub-tours from both the parents so as to construct a new tour of TSP. The method is evaluated from the viewpoint of tour quality and CPU time for ten well-known benchmarks e.g. dj38, qa194 ... ch71009.tsp in the TSP website of Georgia Institute of Technology. We compare SRX with the conventional crossover operators; variant of the Maximal Preservative Crossover operator (MPX3), variant of the Greedy Sub-tour Crossover operator (GSX2) and variant of the Edge Recombination Crossover operator (ERX6), and show that the SRX succeeded in finding better solution and running faster than the conventional methods. © 2010 ISAROB.
-
Kuroda M., Yamamori K., Munetomo M., Yasunaga M., Yoshihara I.
2010 IEEE World Congress on Computational Intelligence, WCCI 2010 - 2010 IEEE Congress on Evolutionary Computation, CEC 2010 2010.12
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：2010 IEEE World Congress on Computational Intelligence, WCCI 2010 - 2010 IEEE Congress on Evolutionary Computation, CEC 2010
This paper proposes a novel crossover operator for solving large-scale traveling salesman problems (TSPs) by using a Hybrid Genetic Algorithm (HGA) with Lin-Kernighan heuristic for local search and we tentatively name Zoning Crossover (Z-Cross). The outline of Z-Cross is firstly to set a zone in the travelling area according to some rules, secondly to cut edges connecting cities between inside and outside the zone, thirdly to exchange edges inside the zone of one parent and those of the other parent, and lastly to reconnect sub-tours and isolated cities, which come about in the 3rd step mentioned above, so as to construct a new tour of TSP. The method is compared with conventional three crossovers; those are the Maximal Preservative Crossover, the Greedy Sub-tour Crossover and the Edge Recombination Crossover, and evaluated from the viewpoints of tour quality and CPU time. Ten benchmarks are selected from the well-known TSP website of Georgia Institute of Technology, whose names are xqf131, xqg237, ..., sra104815. The experiments are performed ten times for each crossover and each benchmark and show that the Z-Cross succeeds in finding better solution and running faster than the conventional methods. Six benchmarks with size from 39,603 to 104,815 cities are selected from the TSP website and challenged the records of tour lengths. The Z-Cross betters the record of the problem rbz43748 and approaches to solutions less than only 0.02% over the known best solutions for five instances. © 2010 IEEE.
-
Quest for genetic information hidden behind disorder in DNA sequences
Koyama Y., Nishimuta K., Yamamori K., Yasunaga M., Yoshihara I.
Proceedings of the 15th International Symposium on Artificial Life and Robotics, AROB 15th'10 824 - 827 2010.12
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：Proceedings of the 15th International Symposium on Artificial Life and Robotics, AROB 15th'10
Most of conventional base sequences are analyzed by order of base sequences, for example, pattern matching. Pattern matching compares unknown base sequences with that of known gene to find similar patterns and to identify gene information. We try to search for hidden information in DNA sequences without pattern matching. We focus on disorder of base sequences, because disorder analysis is available, if we do not know particular function of genes. We use the exponent α of 1/f α fluctuation and self-information as indices of disorder. Our experimental data are ribosomal protein of eukaryotic species. The exponent α is calculated for three kinds of data, i.e. whole base sequences, base sequences in exon or intron. The average of α in exon regions are smaller than that in intron regions. It suggests that exon regions are somewhat more ordered than intron regions. SOM is used to look for similarity of species by self-information which is calculated for codons of base sequences. SOM shows that self-information is usable for a classification of species. © 2010 ISAROB.
-
A Design of Self-defect-compensatable Hardware Neuron for Multi-layer Neural Networks Reviewed
Kunihito YAMAMORI, Keisuke TASHIRO, Masamichi KUSANO, Ikuo YOSHIHARA
Proceedings of 2010 IEEE International Symposium of Defect and Fault Tolerance in VLSI Systems 82 - 89 2010.10
Language：English Publishing type：Research paper (international conference proceedings)
-
A Proposal for Zoning Crossover of Hybrid Genetic Algorihms for Large-scale Travelling Salesman Problems Invited
Masafumi KURODA, Kunihito YAMAMORI, Masaharu MUNETOMO, Morotoshi YASUNAGA, Ikuo YOSHIHARA
Proceeding of the 2010 IEEE World Congress on Computational Intelligence, IEEE CEC 2010 646 - 651 2010.7
Language：English Publishing type：Research paper (international conference proceedings)
-
A New Method to Estimate Food's Function by Genetic Programming
Mayumi KAMIGUCHI, Kunihito YAMAMORI, Ikuo YOSHIHARA, Kazuo NISHIYAMA, Kiyoko NAGAHAMA
1 - 8 2010.3
Language：Japanese Publishing type：Research paper (other academic)
-
Parallel Genetic Programming Model with Remote Memory Access and Multi-threading
Shingo KUROSE, Mayumi KAMIGUCHI, Kunihito YAMAMORI, Masaru AIKAWA, Ikuo YOSHIHARA
1 - 8 2010.3
Language：Japanese Publishing type：Research paper (other academic)
-
Quest for Genetic Information behind Disorder in DNA Sequences Reviewed
Yuka KOYAMA, Kentaro NISHIMUTA, Kunihito YAMAMORI, Moritoshi YASUNAGA, Ikuo YOSHIHARA
Proceedings of the fifteenth International Symposium on Artificial Life and Robotics 828 - 831 2010.2
Language：English Publishing type：Research paper (international conference proceedings)
-
Development of A Novel Crossover of Hybrid Genetic Algorithms for Large-scale Traveling Salesman Problems Reviewed
Masafuni KURODA, Kunihito YAMAMORI, Masaharu MUNETOMO, Morotoshi YASUNAGA, Ikuo YOSHIHARA
Proceedings of the fifteenth International Symposium on Artificial Life and Robotics 832 - 835 2010.2
Language：English Publishing type：Research paper (international conference proceedings)
-
Kuno T., Kamiguchi M., Yamamori K., Yoshihara I., Nagahama K.
Proceedings of the 14th International Symposium on Artificial Life and Robotics, AROB 14th'09 630 - 633 2009.12
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：Proceedings of the 14th International Symposium on Artificial Life and Robotics, AROB 14th'09
We developed a system to estimate physiological activities of foods from protein expression levels using artificial neural networks (ANNs). Since protein expression levels and physiological activities are measured in multiple times for a constituent, we employ a simple regression analysis to find appropriate correspondence between physiological activities and protein expression levels. The range of physiological activities are from 0 to Z (Z > 1), they cannot directly use training signals of ANNs because the output of a neuron is limited from zero to one. To tackle this problem, we introduce two parameters K and l to the activation function of our system like as (f(x) = K/1+e -1x ). Our system is based on three-layer ANN and back-propagation algorithm is employed as training algorithm. Experimental results showed that our system can estimate more accurate than that of ANNs with normalized training samples for antioxidant stress activity. ©ISAROB 2009.
-
An Automatic Model Building for Screening Functional Foods with GP Reviewed
Kunihito Yamamori, Ikuo Yoshihara, Mayumi Kamiguchi, Kazuo Nishiyama
ICROS-SICE International Joint Conference 2009 3679 - 3684 2009.8
Language：English Publishing type：Research paper (international conference proceedings)
-
Estimation of Physiological Activity Values from Protein Expression Levels using GP
Mayumi Kamiguchi, Kunihito Yamamori, Ikuo Yoshihara, Kiyoko Nagahama
2009 ( 25 ) 21 - 24 2009.3
Language：Japanese Publishing type：Research paper (other academic)
-
Feature extraction of protein expression levels based on classification of functional foods with SOM Reviewed
Tamon Fukushima, Kunihito Yamamori, Ikuo Yoshihara, Kiyoko Nagahama
Artificial Life and Robotics 13 ( 2 ) 543 - 546 2009.3
Language：English Publishing type：Research paper (scientific journal)
-
Estimation of Physiological Activity Values from Protein Expression Levels using GP
KAMIGUCHI Mayumi, YAMAMORI Kunihito, YOSHIHARA Ikuo, NAGAHAMA Kiyoko
IPSJ SIG technical reports 2009 ( 25 ) 21 - 24 2009.2
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：Information Processing Society of Japan (IPSJ)
This research reports a new procedure to estimate physiological activity values of foods from protein expression levels using Genetic Programming (GP). Protein expression levels and physiological activities by refined constituents of foods are mesured in advance and they are used to make a model by GP. Then physiological activity values of food extracts are estimated by the above model, from their protein expression levels only.
-
Development of Physiological Activity Estimation Method of Foods Using Amplitude Extended Neural Networks Reviewed
T.Kuno, M.Kamiguchi, K.Yamamori, I.Yoshihara, K.Nagahama
Proceedings of The Fourteenth International Symposium on ARTIFICIAL LIFE AND ROBOTICS 658 - 661 2009.2
Language：English Publishing type：Research paper (international conference proceedings)