===============================================
Technical Committee on Infomation-Based Induction Sciences and Machine Learning (IBISML)
Chair: Naonori Ueda (NTT) Vice Chair: Takashi Washio (Osaka Univ.), Tomoyuki Higuchi (ISM)
Secretary: Masashi Sugiyama (Tokyo Inst. of Tech.), Tsuyoshi Ide (IBM)
Assistant: Daichi Mochihashi (ISM), Shigeyuki Ohba (Kyoto Univ.)
DATE:
Mon, Mar 4, 2013 14:15 - 17:00
Tue, Mar 5, 2013 09:55 - 17:00
PLACE:
(Dr. Ichiro Takeuchi)
TOPICS:
----------------------------------------
Mon, Mar 4 PM (14:15 - 15:30)
----------------------------------------
(1) 14:15 - 14:40
Passive-Aggressive Algorithm with Bias
Daigo Tateishi, Kohei Hatano, Eiji Takimoto (Kyushu Univ.)
(2) 14:40 - 15:05
On Variational Free Energy of Normal Mixtures with General Dirichlet Distribution
Fumito Nakamura, Sumio Watanabe (Tokyo Inst. of Tech.)
(3) 15:05 - 15:30
Model-Based Policy Gradients with Parameter-Based Exploration by Least-Squares Conditional Density Estimation
Syogo Mori, Voot Tangkaratt, Tingting Zhao (Tokyo Inst. of Tech.), Jun Morimoto (ATR), Masashi Sugiyama (Tokyo Inst. of Tech.)
----- Break ( 15 min. ) -----
----------------------------------------
Mon, Mar 4 PM (15:45 - 17:00)
----------------------------------------
(4) 15:45 - 16:10
Task graphs for machine learning algorithms
Sayaka Akioka, Yoichi Muraoka, Hayato Yamana (Waseda Univ.)
(5) 16:10 - 16:35
An Application of Minimum Description Length Criterion to the Model Selection of Discrete Mixture Models
Yasuaki Akazawa, Masato Inoue (Waseda Univ.)
(6) 16:35 - 17:00
Dual Averaging and Proximal Gradient Descent for Online Alternating Direction Multiplier Method
Taiji Suzuki (Univ. of Tokyo)
----------------------------------------
Tue, Mar 5 AM (09:55 - 10:45)
----------------------------------------
(7) 09:55 - 10:20
A Privacy Preserving Distributed Calculation Method of Maximum A Posteriori Estimator for Linear Regression Models
Akito Nakai, Tota Suko, Toshiyasu Matsushima (Waseda Univ.)
(8) 10:20 - 10:45
An evaluation of the performance on decoding spread spectrum digital watermarking against collusion attack.
Kouji Yamauchi, Masato Inoue (Waseda Univ.)
----- Break ( 15 min. ) -----
----------------------------------------
Tue, Mar 5 AM (11:00 - 11:50)
----------------------------------------
(9) 11:00 - 11:25
Non-Achievability of Asymptotic Minimax Regret without Knowledge of the Sample Size
Kazuho Watanabe (NAIST), Teemu Roos, Petri Myllymaki (Helsinki Inst. for Information Tech.)
(10) 11:25 - 11:50
Elimination of redundant features by Kernel Methods and Random Matrix Theory
Hideko Kawakubo, Hiroaki Yoshida (Ochanomizu Univ)
----- Break -----
----------------------------------------
Tue, Mar 5 PM (13:00 - 14:00)
----------------------------------------
(11) 13:00 - 14:00
[Invited Talk]
Meta-heuristics
Mutsunori Yagiura (Nagoya Univ.)
----- Break ( 10 min. ) -----
----------------------------------------
Tue, Mar 5 PM (14:10 - 15:25)
----------------------------------------
(12) 14:10 - 14:35
Constrained Least-Squares Density-Difference Estimation
Tuan Duong Nguyen, Marthinus Christoffel du Plessis (Tokyo Inst. of Tech.), Takafumi Kanamori (Nagoya Univ.), Masashi Sugiyama (Tokyo Inst. of Tech.)
(13) 14:35 - 15:00
*
Yusuke Kishi, Takuma Nakamura, Tatsuhiro Harada, Takashi Matsumoto (Waseda Univ.)
(14) 15:00 - 15:25
Evaluation of the limit performance of single frame super-resolution
Masanori Kawakita, Kouratarou Yamaguchi, Norikazu Takahashi, Jun'ichi Takeuchi (Kyushu Univ.)
----- Break ( 15 min. ) -----
----------------------------------------
Tue, Mar 5 PM (15:40 - 17:00)
----------------------------------------
(15) 15:40 - 16:05
Hierarchical Multi-label Classification on Statistical Decision Theory
Kiyohito Yamamoto, Tota Suko, Toshiyasu Matsushima (Waseda Univ.)
(16) 16:05 - 16:30
Computing pathwise SVMs by Using Non-Support Vector Screening
Kohei Ogawa, Yoshiki Suzuki, Ichiro Takeuchi (Nagoya Inst. of Tech.)
(17) 16:30 - 16:55
Similarity adaptation for label propagation based on local linear reconstruction
Masayuki Karasuyama, Hiroshi Mamitsuka (Kyoto Univ.)
(18) 16:55 - 17:00
# Information for speakers
General Talk will have 20 minutes for presentation and 5 minutes for discussion.
=== Technical Committee on Infomation-Based Induction Sciences and Machine Learning (IBISML) ===
Last modified: 2013-03-05 03:15:26
|
Notification: Mail addresses are partially hidden against SPAM.
|