Copyright (C) IBIS. 1999.
All rights reserved.
First drafted 1999.
Last revised
Mon August 23 15:12 JST 1999.
1999 Workshop on Information-Based Induction Sciences Program
August 26th
- Opening Remarks 12:20-12:30
- IBIS'99 Committee of Planning Directors
- Session 1
(Chair : Yoshiyuki Kabashima) 12:30--14:30
- Fumiyasu Komaki (Univ. of Tokyo) 12:30--13:20
``Theory and Application of Predictive Distributions''(Invited Speech)
- Ritei Shibata (Keio Univ.) 13:20--14:10
``Analyzing Financial Time Series Data;
Gap Between Theory and Practice''(Invited Speech)
- Mitsunori Iino (Yasuda Pension Research Inst.),Tohru Ozaki (Inst. Stat. Math.) 14:10-14:30
``Modeling Financial Time Series Using State Space Models''
- Break : 14:30--14:50
- Session 2
(Chair : Kenji Yamanishi) 14:50--16:40
- Manabu Honda (Okazaki National Research Inst.) 14:50-15:40
``Application of Independent Component Analysis to Functional Neuroimaging Data''(Invited Speech)
- Noboru Murata (RIKEN) 15:40--16:00
``Properties of the Empirical Characteristic Function and Its Application to Test for Independence and Independent Component Analysis''
- Shotaro Akaho , Shinji Umeyama (ETL) 16:00--16:20
``Extension of Independent Component Analysis to Canonical Correlation Analysis''
- Kenji Okajima, Hitoshi Imaoka (NEC) 16:20--16:40
``Receptive Fields of Visual Cortical Neurons as Derived by Infomax''
- Break: 16:40--17:00
- Session 3
(Chair : Toshiyasu Matsushima) 17:00-19:10
- Mary Inaba, Hiroshi Imai, Kunihiko Sadakane (Univ. of Tokyo) 17:00--17:50
``Voronoi Diagram and Clustering in Information Geometry
-- Their Computational and Combinatorial Complexity'' (Invited Speech)
- Takafumi Kanamori (The Graduate Univ. for Advanced Studies),Hidetoshi Shimodaira (Inst. Stat. Math.) 17:50--18:10
``Optimization of Active Learning Using an Information Criterion for the Maximum Weighted Log-likelihood Estimator''
- Masahito Hayashi (Kyoto Univ.) 18:10--18:30
``Large Deviation Evaluation in Adaptive Measurement''
- Tsutomu Kawabata (Univ. of Electro-Commun.) 18:30--18:50
``Variable Length Code with Generalized Cost Functions''
- Kouki Hojo (Osaka Univ.),B.Y.Ryabko (Siberian State Academy of Telecom.), Joe Suzuki (Osaka Univ.) 18:50--19:10
"If Stationary Ergodicity Is Assumed, Essence Will Be Missed (1)
--Comparing Kolmogorov Complexity and Shannon Entropy--"
- Banquet 19:30--21:30
- Poster Session 21:30--22:30
- Hiroshi Yamakawa, Hiroyuki Okada, Nobuo Watanabe (Fujitsu)
``Data Decomposition Algorithm That Extracts Subsets of Data Each of Which Contains a Rule''
- Chikara Matsunaga (FOR-A), Kenichi Kanatani (Gunma Univ.)
``Stabilizing Moving Camera Calibration from Images by the Geometric AIC''
- Yoichi Motomura (ETL), Nikos Vlassis, Ben Krose (Univ. of Amsterdam)
``Mutual Information Based Criterion for Feature Selection of Robot Localization''
- Masashi Sugiyama, Hidemitsu Ogawa (Tokyo Inst. of Tech.)
"Functional Analytic Approach to Model Selection
--Subspace Information Criterion"
- Yasunari Maeda (NTT)
``A Note on Positioning Unknown Words in an Existing Thesaurus Based upon Statistical Decision Theory''
- Kohtaro Tadaki (Hokkaido Univ.)
``Algorithmic Information Theory and Fractal Sets''
- Katsuyuki Hagiwara , Kazuhiro Kuno (Mie Univ.)
``A Geometrical Interpretation of the Gradient Descent Learning of Multilayer Perceptron''
- M.Asano, Y.Nakada, Y.Nakajima, T.Matsumoto (Waseda Univ.)
"Quadratic Approximation vs. MCMC for Hierarchical Bayesian Neural Nets :
Model Comparisons and Predictive Mean Computations "
- Yoichi Seki (Gunma Univ.), Takako Tsutsui (National Inst. of Public Health), Takaya Miyano(Hirosaki Univ.)
``Model Selection of Tree Regression Models Including Linear Regression Terms''
- Tatsuto Murayama, Yoshiyuki Kabashima (Tokyo Inst. of Tech.), David Saad (Aston Univ.)
``Statistical Mechanics of Gallager-type Error-Correcting Codes''
- Maomi Ueno (Chiba Univ.)
``Asymptotic Analysis of Log-Likelihood for Bayesian Networks''
August 27th
- Breakfast 7:30-- 8:30
- Special Session
(Chair : Joe Suzuki) 8:30-12:30
- Commentators :
Jorma Rissanen(IBM), Te Sun HAN (Univ. of Electro-Commun.), Ritei Shibata (Keio Univ.), Kenji Yamanishi (NEC)
- Jorma Rissanen (IBM) 8:30--9:40
"Information in Data, Universal Modeling, and Generalized Minimax Bounds" (Invited Speech)
- Part 1 (consistency): 9:40--10:25
- Taichi Hayasaka (Toyohashi Univ. of Tech.),Katsuyuki Hagiwara (Mie Univ.), Naohiro Toda (Aichi Prefectural Univ.), Shiro Usui (Toyohashi Univ. of Tech.)
"Statistical Model Selection for the Nonlinear Regression Models with Orthogonal Basis Functions"
- Joe Suzuki (Osaka Univ.)
"On the Error Probability of Model Selection for Classification"
- Break : 10:25--10:40
- Part 2 (stochastic complexity): 10:40--11:25
- Kenji Yamanishi (NEC)
"Extended Stochastic Complexity in Individual Sequence Analysis"
- Masayuki Gotoh, Toshiyasu Matsushima, Shigeichi Hirasawa
(Waseda Univ.)
"On Asymptotic Normality of Density Functions and Its Applications"
- Part 3 (alternative approaches): 11:25--12:30
- Ritei Shibata, Shigeo Kamitsuji (Keio Univ.)
"Likelihood of a Stochastic Neural Network and Its Model Selection"
- Genshiro Kitagawa (Inst. Stat. Math.), Sadanori Konishi (Kyushu Univ.)
"Information Criteria--AIC,GIC and EIC"
- Kenichi Kanatani (Gunma Univ.),Yasushi Kanazawa (Toyohashi Univ. of Tech.)
"Stabilizing Image Mosaicing by the Geometric AIC"
- Session 4
(Chair : Noboru Murata) 13:30--15:20
- Shiro Ikeda (JST) 13:30--14:20
``Two Related Subjects of the EM Algorithm
-- The Wake-Sleep Algorithm and the Recursive EM Algorithm --''
(Invited Speech)
- Seiichi Shimasaki, Yoshiyuki Kabashima
(Tokyo Inst. of Tech.) 14:20--14:40
``Dynamical Analysis of the Wake-Sleep Algorithm''
- T.Matsumoto, J.Sugi, M.Saito, Y.Nakajima, and H.Hamagishi
(Waseda Univ.) 14:40--15:00
"Selection of Model Dynamical Systems for Nonlinear Time Series Prediction Problems with Hierarchical Bayesian Neural Nets"
- Yasuaki Oishi (Univ. of Tokyo) 15:00--15:20
``System Identification and Learning Theory for Robust Control ''
- Break : 15:20--15:35
- Session 5
(Chair : Junichi Takeuchi) 15:35--17:05
- Shinto Eguchi (Inst. Stat. Math.) 15:35--16:25
``Common Room with Statisticians and Neural Networkers
-- Near-Parametric Approach to Probabilistic Modeling -- ''(Invited Speech)
- Kenji Fukumizu (Riken) 16:25--16:45
``Generalization Error of the Maximum Likelihood Estimation of Layered Models''
- Takafumi Mukouchi (NTT) 16:45--17:05
``Generalization Error Bound by Margin Distribution''
- Break : 17:05--17:20
- Session 6
(Chair : Hirosuke Yamamoto) 17:20--18:20
- Hirotoshi Taira (NTT), Masahiko Haruno (ATR) 17:20--17:40
``SVM Text Categorization''
- Hang Li, Kenji Yamanishi (NEC) 17:40--18:00
``Text Classification Based on ESC''
- Taisei Kaizoji (International Christian Univ.) 18:00--18:20
``Learning Control of Speculative Chaos''
- Concluding Remarks 18:20-18:30
- IBIS'99 Committee of Planning Directors