Committee |
Date Time |
Place |
Paper Title / Authors |
Abstract |
Paper # |
IBISML |
2015-11-26 15:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
Robustification of Learning Algorithms using Hinge-loss Takafumi Kanamori (Nagoya Univ.), Shuhei Fujiwara (TopGate), Akiko Takeda (Univ. of Tokyo) IBISML2015-71 |
We propose a unified formation of robust learning methods for classification and regression problems.
In the learnin... [more] |
IBISML2015-71 pp.139-146 |
IBISML |
2015-11-26 15:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
Learning Structure of Partial Markov Random Field via Partitioned Ratio Song Liu (ISM), Taiji Suzuki (Tokyo Tech.), Masashi Sugiyama (UTokyo), Kenji Fukumizu (ISM) IBISML2015-72 |
A new concept, partitioned ratio is proposed to find the partial connectivity of the Markov random field. First we argue... [more] |
IBISML2015-72 pp.147-151 |
IBISML |
2015-11-26 15:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
Full-span log-linear model with L1 reguralization and its performance Kazuya Takabatake, Shotaro Akaho (AIST) IBISML2015-73 |
The full-span log-linear model is a log-linear model that has sufficient number of basis functions so that it is able to... [more] |
IBISML2015-73 pp.153-157 |
IBISML |
2015-11-27 14:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
Minimum required data amount in Bayesian inference from the viewpoint of specific heat Satoru Tokuda, Kenji Nagata, Masato Okada (Univ. of Tokyo) IBISML2015-74 |
The accuracy of Bayesian inference depends on the number of samples or noise. Sample size or noise level often changes t... [more] |
IBISML2015-74 pp.159-166 |
IBISML |
2015-11-27 14:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
Extended KL-UCB Policy for the Budgeted Multi-Armed Bandits Ryo Watanabe, Atsuyoshi Nakamura, Mineichi Kudo (Hokkaido Univ.) IBISML2015-75 |
[more] |
IBISML2015-75 pp.167-174 |
IBISML |
2015-11-27 14:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
Stochastic process representation of R-tree Masahiro Nakano, Xiaomeng Wu, Minoru Mori, Akisato Kimura, Kunio Kashino (NTT) IBISML2015-76 |
This paper discusses a new stochastic process that represents the innitely exchangeable rectangular partitioning of a m... [more] |
IBISML2015-76 pp.175-182 |
IBISML |
2015-11-27 14:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
Recursive Ensemble Land Cover Classification for Few Training Data and Many Class Yu Oya, Katsutoshi Kanamori, Hayato Ohwada (TUS) IBISML2015-77 |
Many global and environmental applications require land use and land cover information. A land cover classification is o... [more] |
IBISML2015-77 pp.183-188 |
IBISML |
2015-11-27 14:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
An MLE Algorithm for Stochastic Strictly Piecewise Languages and its Implications Chihiro Shibata (TUT) |
[more] |
|
IBISML |
2015-11-27 14:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
Energy Disaggregation Based on Weighted Semi-Binary NMF Masako Matsumoto, Yu Fujimoto, Yasuhiro Hayashi (Waseda Univ.) IBISML2015-78 |
The authors propose an application of weighted semi-binary non-negative matrix factorization for the energy disaggregati... [more] |
IBISML2015-78 pp.189-194 |
IBISML |
2015-11-27 14:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
Clustering Features based on Simultaneous Manifold Learning Masayuki Karasuyama (NIT), Hiroshi Mamitsuka (Kyoto Univ.) IBISML2015-79 |
[more] |
IBISML2015-79 pp.195-201 |
IBISML |
2015-11-27 14:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
Nonnegative matrix factorization of test scores Atsunori Kanemura (AIST), Hiroko Onari, Manabu Shikauchi, Masataka Hashimoto (Recruit Career), Shotaro Akaho (AIST) IBISML2015-80 |
When assessing skills of humans or designing itemsets, having fewer number of items is cost-efficient both in time and m... [more] |
IBISML2015-80 pp.203-208 |
IBISML |
2015-11-27 14:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
Multi-label Classification Using Logistic Loss Matrix Factorization Keigo Kimura, Mineichi Kudo (Hokkaido Univ.) |
[more] |
|
IBISML |
2015-11-27 14:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
Sufficient Dimension Reduction via Direct Estimation of the Gradients of Logarithmic Conditional Densities Hiroaki Sasaki, Voot Tangkaratt, Masashi Sugiyama (UTokyo) IBISML2015-81 |
[more] |
IBISML2015-81 pp.209-216 |
IBISML |
2015-11-27 14:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
Non-Gaussian Component Analysis with Log-Density-Gradient Estimation Hiroaki Sasaki, Gang Niu, Masashi Sugiyama (UTokyo) IBISML2015-82 |
[more] |
IBISML2015-82 pp.217-224 |
IBISML |
2015-11-27 14:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
An Extension of Modularity to Pseudo Clique for Overlapping Communities Sadamori Koujaku, Mineichi Kudo (Hokkaido Univ) IBISML2015-83 |
[more] |
IBISML2015-83 pp.225-232 |
IBISML |
2015-11-27 14:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
Evaluation of Hidden Semi-Markov Models with Interval State Node Hiromi Narimatsu, Hiroyuki Kasai (UEC) IBISML2015-84 |
This paper presents an extended HMM that models both sequence length and sequence interval of sequential data. The struc... [more] |
IBISML2015-84 pp.233-240 |
IBISML |
2015-11-27 14:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
Performance degradation of AMP for Ising perceptron when the system size is small Arise Kuriya, Toshiyuki Tanaka (Kyoto Univ.) IBISML2015-85 |
Approximate Massage Passing (AMP) algorithm, proposed by Donoho et al., is derived from Belief Propagation (BP) algorith... [more] |
IBISML2015-85 pp.241-247 |
IBISML |
2015-11-27 14:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
Secure Approximation Guarantee for Private Empirical Risk Minimization with Homomorphic Encryption Toshiyuki Takada, Hiroyuki Hanada (NIT), Jun Sakuma (Univ.Tsukuba), Ichiro takeuchi (NIT) IBISML2015-86 |
Privacy concern has been increasingly important in many machine learning problems. In this paper, we study empirical ris... [more] |
IBISML2015-86 pp.249-256 |
IBISML |
2015-11-27 14:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
Regularization by local distributional smoothing Takeru Miyato, Shin-ichi Maeda, Masanori Koyama, Ken Nakae, Shin Ishii (Kyoto Univ.) IBISML2015-87 |
Smoothness regularization is a popular method to decrease generalization error. We propose a novel regularization techni... [more] |
IBISML2015-87 pp.257-264 |
IBISML |
2015-11-27 14:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
Bayesian Masking for Sparse Feature Selection Yohei Kondo (Kyoto Univ.), Kohei Hayashi (NII), Shin-ichi Maeda (Kyoto Univ.) IBISML2015-88 |
In linear regression, we can reduce the weights of irrelevant features by L2 or L1 regularization. However, such a regul... [more] |
IBISML2015-88 pp.265-272 |