Japanese (日本語)

Susumu Katayama

School of Engineering, Univerisity of Miyazaki
1-1 W. Gakuen Kibana Dai, Miyazaki, Miyazaki 889-2155, Japan
email: skata@cs.miyazaki-u.ac.jp
Tel/Fax: +81-985-58-7941

Personal history

Education

Bachelor in Agriculture from University of Tokyo in 1995,
Master in Engineering from University of Tokyo in 1997,
Ph.D. in Engineering from Tokyo Institute of Technology in 2000.

Jobs

2000-2001 Research Associate at Tokyo Institute of Technology;
2001-2002 Postdoctoral Scientist at NEC Research Institute, Princeton;
2002- Research Associate/Assistant Professor at University of Miyazaki.

MagicHaskeller on the Web

Selected papers

Artificial General Intelligence

S. Katayama (2019):
Computable Variants of AIXI which are More Powerful than AIXItl ,
Journal of Artificial General Intelligence , Vol. 10, No. 1, 1-23. (Kudos)

Inductive Functional Programming

(Also try the above box and/or see the MagicHaskeller page for trying the software.)

S. Katayama (2012):
An Analytical Inductive Functional Programming System that Avoids Unintended Programs ,
Proceedings of the ACM SIGPLAN 2012 Workshop on Partial Evaluation and Program Manipulation, 43-52.

S. Katayama (2008):
Efficient Exhaustive Generation of Functional Programs using Monte-Carlo Search with Iterative Deepening ,
PRICAI 2008: Trends in Artificial Intelligence, Proceedings of 10^th Pacific Rim International Conference on Artificial Intelligence, LNAI 5351, Springer Verlag, 199-211.

S. Katayama (2005):
Systematic Search for Lambda Expressions,
Sixth Symposium on Trends in Functional Programming (TFP2005) (informal) Proceedings. 195-205
Updated formal version is in preparation for publishing. Please contact me personally in order to obtain a copy, because I do not know how the copyright agreement will be.

S. Katayama (2004):
Power of Brute-force Search in Strongly-Typed Inductive Functional Programming Automation,
PRICAI 2004: Trends in Artificial Intelligence, Proceedings of 8^th Pacific Rim International Conference on Artificial Intelligence, Springer LNAI 3157. 75-84
Abstract

Reinforcement Learning

S. Katayama, H. Kimura, and S. Kobayashi (2000):
A Universal Generalization for Temporal-Difference Learning using Haar Basis Functions,
Proceedings of the 17th International Conference on Machine Learning. Morgan Kaufmann.

S. Katayama and S. Kobayashi (1999):
Logarithmic-time Updating Algorithm for TD(lambda) Learning,
Journal of Japanese Society for Artificial Intelligence, 14, 119-130 (in Japanese).
(English translation is available here.) C++ codes (FiniteTD.C, FiniteTD.H)

Tools

prof2xdu -- short program for visualizing time and allocation profiling results of Haskell programs

This program converts a .prof file into the format that the xdu command understands, and invokes xdu.

(So requires xdu.)

ps2xdu -- short program for visualizing results of ps command for identifying which programs (processes and their descendants) use lots of CPU time and memory resource.

This program converts results of ps command into the format that the xdu command understands, and invokes xdu.