Message
My research and that of my graduate students focus on evolutionary computation, artificial life, and complex adaptive systems. I want to bring up talented people with the creatively. My hobby is to play tennis, traveling, and gardening.

>PAGE TOP

Publications(January 2001 - December 2001)

1. Y. Sato, "Proposal for a Field-Evolvable Hardware based on a Microprocessor Incorporated Flash Memory", Proc. of the 2001 Congress on Evolutionary Computation, IEEE Press, pp. 608-615, Seoul, Korea, May 2001.
   Abstract - A new idea for evolvable hardware based on a microprocessor is proposed. Evolvable hardware is a new direction in hardware research that fuses evolutionary computation and reconfigurable logic LSI circuits. In recent years, there has been much research using Programmable Logic Devices (PLD) and Field Programmable Gate Arrays (FPGA). In particular, the application of digital circuit evolution to engineering fields has already begun. On the other hand, long learning time, difficulty to predict when an effective capability will appear, large chip size and other such problems have hindered progress in diffusion into engineering fields. Here, we propose register transfer level evolution performed on a microprocessor as a means of addressing these problems. Specifically, we propose (1) incorporating flash memory into the microprocessor to allow on-board programming and reprogramming, (2) using genetic algorithms to provide a register transfer level learning capability, and (3) the use of a framework that provides for the coexistence of static programs and programs that self-organize through learning. On the basis of a simple hand-design, we concluded that the proposed method is more effective in terms of learning efficiency and reliability than the conventional approach using FPGA and PLD.
2. Y. Sato, "Proposal for a Register Transfer Level Evolution on a Microprocessor Incorporated Flash Memory", Proc. of the 2001 Genetic and Evolutionary Computation Conference, Morgan Kaufmann Publishers, p. 1216, San Francisco, USA, July 2001.
   Abstract - A new idea for evolvable hardware based on a microprocessor is proposed. In recent years, there has been much research using Programmable Logic Devices (PLD) and Field Programmable Gate Arrays (FPGA). In particular, the application of digital circuit evolution to engineering fields has already begun. On the other hand, long learning time, difficulty to predict when an effective capability will appear, large chip size and other such problems have hindered progress in diffusion into engineering fields. Here, we propose register transfer level evolution performed on a microprocessor as a means of addressing these problems.
3. R. Goto and Y. Sato, "Motion Analysis of Moving Objects with Genetic Algorithms", Proc. of the 2001 Genetic and Evolutionary Computation Conference, Morgan Kaufmann Publishers, pp. 1276-1283, San Francisco, USA, July 2001.
   Abstract - A series of basic experiments were conducted to investigate the applicability of genetic algorithms (GAs) to complex problems including (1) multiobjective optimization, (2) time-series prediction, (3) analysis from noisy observational data, (4) implicit function solutions, and other problems. More specifically, this work addresses the problem of tracking a moving object as it tries to camouflage its movements (say to avoid being pursued or attacked) by exploring the effectiveness of genetic algorithms in predicting the future position of moving objects from noisy time-series data alone such as obtained from radar or observational equipment. For the present purpose, we assume a moving object such as a ship that would be unlikely to maintain forward momentum with frequent sharp course corrections. We therefore assume a moving object that attempts to camouflage its movements by pursuing sine-wave, zigzag, or linear motion. The experimental findings demonstrate that the genetic algorithm-based approach yields more accurate solution than the conventional approach based on regression analysis and assuming uniform velocity linear motion.
4. Y. Sato, M. Kawamoto, and M. Chizaki, "Verification of Text Transcription History by using Evolutionary Algorithms", Advances in Artificial Life, Lecture Notes in Artificial Intelligence 2159 (Proc. of the 6th European Conference on Artificial Life), J. Kelemen and P. Sosik (Eds.), Springer-Verlag, pp. 444-453, Prague, Czech Republic, September 2001.
   Abstract - The application of ALife methods to the field of historical document genealogy is reported. At first, the conventional analytical approaches to Greek and Latin text transcription history are described and the problems associated with them are discussed. Next, focusing on the similarities between text transcription history and the Tierra system, we propose the use of evolutionary algorithms for the verification of text transcription genealogies with the purposes of breaking through the limitations of conventional analysis of Greek and Latin text transcription history and increasing the reliability of those methods. In this report, as the first step, we deal only with the mutations involved in copying, and attempt to make use of them in the verification of genealogy. We conducte computer simulation experiments based on existing manuscript data and demonstrate the feasibility of effectively using evolutionary algorithms as one means of verifying the currently proposed Greek and Latin text genealogies.
5. Y. Sato, "Interactive Evolution of Adaptive Parameter for Speaker Verification Systems", Proc. of the IPSJ Symposium Series Vol. 2001, No. 12, pp. 227-232, Kyoto, Japan, October 2001. (in Japanese)
   Abstract - The application of interactive evolutionary algorithms (EA) is proposed as a means of improving the accuracy of a speaker verification system through the adaptation of semi-continuous hidden Markov models (HMM). Interactive EA is an evolutionary algorithm in which the genetic operator evolves on the basis of the user's subjective criteria as the result of interaction between the user and a computer. The use of interactive EA is spreading, primarily in artistic fields such as computer graphics and music. We employed an interactive EA to search for the parameter to create a speaker HMM through speaker adaptation of a speaker-independent phoneme HMM that was obtained earlier through trial and error experimentation by other researchers. Specifically, the parameter is the confidence coefficient of a maximum a posteriori (MAP) probability estimation. Our motivation was the belief that it would be possible to make good use of the special feature of evolutionary algorithms that they can be applied effectively to problems for which the nature of the target function is not well known. The result was an observed speaker verification accuracy of 6% for voice recorded by telephone and 1% for voice recorded from a microphone.
6. Y. Sato, "Proposal for a Field-Evolvable Hardware based on a Microprocessor Incorporated Flash Memory", Proc. of the IPSJ Symposium Series Vol. 2001, No. 12, pp. 115-122, Kyoto, Japan, October 2001. (in Japanese)
   Abstract - A new idea for evolvable hardware based on a microprocessor is proposed. Evolvable hardware is a new direction in hardware research that fuses evolutionary computation and reconfigurable logic LSI circuits. In recent years, there has been much research using Programmable Logic Devices (PLD) and Field Programmable Gate Arrays (FPGA). In particular, the application of digital circuit evolution to engineering fields has already begun. On the other hand, long learning time, difficulty to predict when an effective capability will appear, large chip size and other such problems have hindered progress in diffusion into engineering fields. Here, we propose register transfer level evolution performed on a microprocessor as a means of addressing these problems. Specifically, we propose (1) incorporating flash memory into the microprocessor to allow on-board programming and reprogramming, (2) using genetic algorithms to provide a register transfer level learning capability, and (3) the use of a framework that provides for the coexistence of static programs and programs that self-organize through learning. On the basis of a simple hand-design, we concluded that the proposed method is more effective in terms of learning efficiency and reliability than the conventional approach using FPGA and PLD.
7. R. Goto and Y. Sato, "Motion Analysis of Moving Objects with Genetic Algorithms", Proc. of the IPSJ Symposium Series Vol. 2001, No. 12, pp. 211-214, Kyoto, Japan, October 2001. (in Japanese)
   Abstract - A series of basic experiments were conducted to investigate the applicability of genetic algorithms (GAs) to complex problems including (1) multiobjective optimization, (2) time-series prediction, (3) analysis from noisy observational data, (4) implicit function solutions, and other problems. More specifically, this work addresses the problem of tracking a moving object as it tries to camouflage its movements (say to avoid being pursued or attacked) by exploring the effectiveness of genetic algorithms in predicting the future position of moving objects from noisy time-series data alone such as obtained from radar or observational equipment. For the present purpose, we assume a moving object such as a ship that would be unlikely to maintain forward momentum with frequent sharp course corrections. We therefore assume a moving object that attempts to camouflage its movements by pursuing sine-wave, zigzag, or linear motion. The experimental findings demonstrate that the genetic algorithm-based approach yields more accurate solution than the conventional approach based on regression analysis and assuming uniform velocity linear motion.
8. Y. Sato, M. Kawamoto, and M. Chizaki, "Verification of Text Transcription History by using Evolutionary Algorithms", Proc. of the IPSJ Symposium Series Vol. 2001, No. 12, pp. 3-10, Kyoto, Japan, October 2001. (in Japanese)
   Abstract - The application of ALife methods to the field of historical document genealogy is reported. At first, the conventional analytical approaches to Greek and Latin text transcription history are described and the problems associated with them are discussed. Next, focusing on the similarities between text transcription history and the Tierra system, we propose the use of evolutionary algorithms for the verification of text transcription genealogies with the purposes of breaking through the limitations of conventional analysis of Greek and Latin text transcription history and increasing the reliability of those methods. In this report, as the first step, we deal only with the mutations involved in copying, and attempt to make use of them in the verification of genealogy. We conducte computer simulation experiments based on existing manuscript data and demonstrate the feasibility of effectively using evolutionary algorithms as one means of verifying the currently proposed Greek and Latin text genealogies.
   Department of computer science

>PAGE TOP