IBIS'98 Program



July 11th (Sat) 13:20-13:30 Opening Remarks by IBIS'98 Committee Session 1 (Chair: Yoshiyuki Kabashima) 13:30-14:20 Inductive and Deductive Inference and Prediction - Learning Models Based on Decision Theory - by Toshiyasu Matsushima(Waseda Univ.) 14:20-15:10 Stochastic Complexity and Jeffreys Mixture Prediction Strategies by Jun-ichi Takeuchi(NEC) 15:10-15:30 Break Session 2 (Chair: Toshiyasu Matsushima) 15:30-16:20 MDL Principle with Application to Image Processing by Shuichi Itoh(Univ.Electro-Comm.), Hideaki Tsuchiya (Univ.Electro-Comm.) 16:20-17:10 Theoretical Accuracy Bounds and Optimal Algorithm in Geometric Estimation by Kenichi Kanatani (Gunma Univ.) 17:10-18:00 Extended Stochastic Complexity and Its Applications to Learning by Kenji Yamanishi (NEC) 18:00-20:30 Banquet+Break Session 3 [Impromptu Session(1)] (Chair: Kenji Yamanishi) 20:30-20:45 Assessing the Uncertainty of Model Selection -A Graphical Approach to Model Selection Diagnosis- by Hidetoshi Shimodaira (Inst. Stat. Math.) 20:45-21:00 Deterministic Annealing EM Algorithm -A Formalization Using Statistical Mechanics Analogy- by Naonori Ueda (NTT) 21:00-21:15 Existence of Overtraining in Batch Learning of Multilayer Neural Networks by Kenji Fukumizu (RIKEN) 21:15-21:30 Block Orthogonal Projection Learning for Multilayer Neural Networks by Seiji Miyoshi (Kobe City Coll. Tech.) 21:30-21:45 Convergence Rate of the Block Orthogonal Projection Algorithm by Kazushi Ikeda (Kyoto Univ.), Seiji Miyoshi (Kobe City Coll. Tech.), Kenji Nakayama (Kanazawa Univ.) 21:45-22:00 Break Session 4 [Impromptu Session(2)] (Chair: Kenji Yamanishi) 22:00-22:15  Bayesian Network that Learns Conditional Probabilities by Neural Networks by Youichi Motomura (ETL) 22:15-22:30  Object Recognition by Composition by Akira Date (Osaka Univ.)  22:30-22:45 A Study of A First Order Asymptotic Theory of Statistical Estimation When the Experimental Scheme Is Controlled by Keiji Matsumoto (Univ. Tokyo) 22:45-23:00 Asymptotic Two-Stage Quantizer by Tsutom Kawabata (Univ.Electro-Comm.) 23:00-23:15 Kolmogorov Complexity of Measure Preserving Transformations by Hayato Takahashi (JST, Tokyo Denki Univ.), Kazuyuki Aihara (Univ. Tokyo)
July 12th (Sun) 7:00-8:00 Breakfast Session 5 (Chair: Jun-ichi Takeuchi) 8:30-9:20 Model Selection and Related Topics in Statistical Sciences by Yukito Iba (Inst.Stat.Math.) 9:20-10:10 Bias of Estimators and Regularization Terms by Noboru Murata (RIKEN) 10:10-10:30 Break Session 6 (Chair: Yukito Iba) 10:30-11:20 Natural Language Analysis Based on MDL Learning by Hang Li(NEC) 11:20-12:10 Japanese Part-of-speech Tagger Using Context Trees by Masahiko Haruno(ATR) 12:10-13:00 Lunch Session 7 (Chair: Joe Suzuki) 13:00-13:50 Using Information Theory To Study the Communicative System and Cognitive Capacities in Ants by Boris Ryabko(Siberian State Academy of Tel. Comput. Science), Zhanna Reznikova (Novosibirsk State Univ.) 13:50-14:40 Independent Component Analysis by Geometrical Learning by Shun-ichi Amari (RIKEN) 14:40-15:00 Break Session 8 (Chair: Noboru Murata) 15:00-15:50 Statistical Mechanics of Learning by Yoshiyuki Kabashima(Tokyo Inst.Tech.) 15:50-16:40 Learning Bayesian Belief Networks Based on the MDL Principle by Joe Suzuki (Osaka Univ.) 16:40-16:50 Concluding Remarks by IBIS'98 Committee/SITA Boad of Planning Directors