Mon, Mar 12 AM Bayesian Models
Chair: Daichi Mochihashi (ISM)
10:00 - 12:15 |
| (1) |
10:00-10:25 |
Chance Adjusted Infinite Relational Model for Asymmetric and Individually Different Relational Data Analysis |
Iku Ohama, Hiromi Iida (Panasonic), Takuya Kida, Hiroki Arimura (Hokkaido Univ.) |
| (2) |
10:25-10:50 |
Bayesian Network Structure Estimation based on the Bayesian/MDL Criteria when both Discrete and Continuous Variables are Present |
Joe Suzuki (Osaka Univ.) |
| (3) |
10:50-11:15 |
Detecting Latent Structural Changes via Latent Dirichlet Allocation |
Masashi Ueda, Ryota Tomioka, Kenji Yamanishi (The Univ. of Tokyo), Katsuhiko Ishiguro, Hiroshi Sawada, Naonori Ueda (NTT) |
| |
11:15-11:25 |
Break ( 10 min. ) |
| (4) |
11:25-11:50 |
Fully Bayesian speaker clustering based on hierarchical structured Dirichlet process mixture model |
Naohiro Tawara, Tetsuji Ogawa (Waseda Univ.), Shinji Watanabe (NTT/MERL), Atsushi Nakamura (NTT), Tetsunori Kobayashi (Waseda Univ.) |
| (5) |
11:50-12:15 |
Hyperparameter Selection of Infinite Gaussian Mixture Model via Widely Applicable Information Criterion |
Takushi Miki, Masahiro Kohjima, Sumio Watanabe (titech) |
Mon, Mar 12 PM Invited Talk (Data mining)
Chair: Tsuyoshi Ide (IBM Japan)
13:30 - 14:30 |
| (6) |
13:30-14:30 |
[Invited Talk]
Subsequence data mining revisited |
Shin Ando (Gunma Univ.) |
| |
14:30-14:40 |
Break ( 10 min. ) |
Mon, Mar 12 PM Reinforcement Learning
Chair: Masashi Sugiyama (Tokyo Institute of Technology)
14:40 - 16:20 |
| (7) |
14:40-15:05 |
Kernel Bellman Equations in POMDPs |
Yu Nishiyama (ISM), Abdeslam Boularias (MPI), Arthur Gretton (UCL), Kenji Fukumizu (ISM) |
| (8) |
15:05-15:30 |
Model Selection of Indirect Value Function Estimation |
Masahiro Kohjima (Tokyo Tech) |
| (9) |
15:30-15:55 |
Apprenticeship Learning for Model Parameters of Partially Observable Environments |
Takaki Makino (Univ. of Tokyo), Johane Takeuchi (HRI-JP) |
| (10) |
15:55-16:20 |
Efficient Sample Reuse in Policy Gradients with Parameter-based Exploration |
Tingting Zhao, Hirotaka Hachiya, Masashi Sugiyama (Tokyo Inst. of Tech.) |
| |
16:20-16:30 |
Break ( 10 min. ) |
Mon, Mar 12 PM Time series analysis
Chair: Shigeyuki Oba (Kyoto University)
16:30 - 17:45 |
| (11) |
16:30-16:55 |
Canonical Multiple Sequence Alignment for High-dimensional Multiple Time-series Analysis |
Takamitsu Matsubara (NAIST/ATR), Jun Morimoto (ATR) |
| (12) |
16:55-17:20 |
Heterogeneous Time Series Clustering based on Infinite-Length Model Distance |
Shunsuke Hirose, Katsuyuki Izumi (SAS) |
| (13) |
17:20-17:45 |
Detecting Long-term Trending Topics in Social Networks |
Shota Saito, Ryota Tomioka, Kenji Yamanishi (The Univ. of Tokyo) |
Tue, Mar 13 AM Statistical Mathematics
Chair: Kei Kobayashi (ISM)
10:00 - 10:50 |
| (14) |
10:00-10:25 |
On d-consistency for high-dimensional discrimination analysis |
Takanori Ayano (Osaka Univ.) |
| (15) |
10:25-10:50 |
An Accuracy Analysis of Bayes Latent Variable Estimation with Redundant Variables |
Keisuke Yamazaki (Tokyo Inst. of Tech.) |
| |
10:50-11:00 |
Break ( 10 min. ) |
Tue, Mar 13 AM Learning Methods
Chair: Kei Kobayashi (ISM)
11:00 - 11:50 |
| (16) |
11:00-11:25 |
A Study on Decision Boundary Stability in Active Learning with Support Vector Machine |
Takeru Takahashi, Yuji Waizumi, Kazuo Hashimoto (Tohoku Univ.) |
| (17) |
11:25-11:50 |
Semi-Supervised Learning of Class Balance under Class-Prior Change by Distribution Matching |
Marthinus Christoffel du Plessis, Masashi Sugiyama (Tokyo Inst. of Tech.) |
Tue, Mar 13 PM Invited Talk (Statistical Mathematics)
Chair: Daichi Mochihashi (ISM)
13:30 - 14:30 |
| (18) |
13:30-14:30 |
[Invited Talk]
Asymptotic estimation theory and Information geometry via computational algebra |
Kei Kobayashi (ISM) |
| |
14:30-14:40 |
Break ( 10 min. ) |
Tue, Mar 13 PM Matrix factorization and applications
Chair: Toshihiro Kamishima (AIST)
14:40 - 15:30 |
| (19) |
14:40-15:05 |
Matrix and Tensor Factorization with Aggregated Observations |
Yoshifumi Aimoto, Hisashi Kashima (Univ. of Tokyo) |
| (20) |
15:05-15:30 |
A Method for Packet Loss Rate Estimation Using Active and Passive Measurements |
Atsushi Miyamoto, Kazuho Watanabe, Kazushi Ikeda (NAIST) |
| |
15:30-15:40 |
Break ( 10 min. ) |
Tue, Mar 13 PM Active learning, Mutual information, ane else
Chair: Daichi Mochihashi (ISM)
15:40 - 17:45 |
| (21) |
15:40-16:05 |
Improving Importance Estimation in Pool-based Batch Active Learning for Approximate Linear Regression |
Nozomi Kurihara, Masashi Sugiyama (Tokyo Inst. of Tech.) |
| (22) |
16:05-16:30 |
Early Stopping Heuristics in Pool-Based Incremental Active Learning for Least-Squares Probabilistic Classifier |
Tsubasa Kobayashi, Masashi Sugiyama (Tokyo Inst. of Tech.) |
| (23) |
16:30-16:55 |
Feature Selection via L1-Penalized Squared-Loss Mutual Information |
Wittawat Jitkrittum, Hirotaka Hachiya, Masashi Sugiyama (Tokyo Inst. of Tech.) |
| (24) |
16:55-17:20 |
Squared-loss Mutual Information Regularization |
Gang Niu, Wittawat Jitkrittum, Hirotaka Hachiya (Tokyo Inst. of Tech.), Bo Dai (Purdue Univ.), Masashi Sugiyama (Tokyo Inst. of Tech.) |
| (25) |
17:20-17:45 |
The Information-Based Inductive Sciences and Biological Learning
-- The Quantum Mechanism of Concept (Digital Linguistics) -- |
Kimiaki Tokumaru (System Engineer) |