Committee |
Date Time |
Place |
Paper Title / Authors |
Abstract |
Paper # |
MBE, NC (Joint) |
2017-12-16 10:55 |
Aichi |
Nagoya University |
New Learning Method Utilizing Singular Regions of RBF Networks Seiya Satoh (AIST), Ryohei Nakano (Chubu Univ.) NC2017-41 |
[more] |
NC2017-41 pp.7-12 |
NC, MBE |
2015-03-16 15:35 |
Tokyo |
Tamagawa University |
Further Speeding Up and Solution Quality Improvement of Singularity Stairs Following Seiya Satoh, Ryohei Nakano (Chubu Univ.) MBE2014-168 NC2014-119 |
In a search space of a multilayer perceptron (MLP), there exists singular regions where any point is I-O equivalent to t... [more] |
MBE2014-168 NC2014-119 pp.289-294 |
NC, MBE (Joint) |
2013-12-21 15:20 |
Gifu |
Gifu University |
Singularity Stairs Following with Limited Numbers of Hidden Units Seiya Satoh, Ryohei Nakano (Chubu Univ.) NC2013-65 |
In a search space of a multilayer perceptron having J hidden units, MLP(J), there exist flat areas called singular regio... [more] |
NC2013-65 pp.69-74 |
NC, NLP |
2013-01-24 09:30 |
Hokkaido |
Hokkaido University Centennial Memory Hall |
Multilayer Perceptron Search Making Good Use of Singular Regions Seiya Satoh, Ryohei Nakano (Chubu Univ.) NLP2012-104 NC2012-94 |
In a search space of multilayer perceptron having J hidden units, MLP(J), there exists a singular flat region created by... [more] |
NLP2012-104 NC2012-94 pp.1-6 |
NC, NLP |
2013-01-24 09:50 |
Hokkaido |
Hokkaido University Centennial Memory Hall |
Multilayer Perceptron Model Selection Using Sampling Utilizing Singularity Stairs Following Takayuki Ohwaki, Ryohei Nakano (Chubu Univ.) NLP2012-105 NC2012-95 |
Multilayer perceptron (MLP) is one of singular statistical models, where it is not guaranteed that any parameter is uniq... [more] |
NLP2012-105 NC2012-95 pp.7-12 |
NC, MBE (Joint) |
2011-12-20 11:20 |
Aichi |
Nagoya Institute of Technology |
Eigen Vector Descent and Line Search for Multilayer Perceptron Seiya Satoh, Ryohei Nakano (Chubu Univ.) NC2011-87 |
As learning methods of a multilayer perceptron (MLP), we have the BP algorithm, Newton's method, quasi-Newton method, an... [more] |
NC2011-87 pp.19-24 |
NC, MBE (Joint) |
2011-12-20 11:45 |
Aichi |
Nagoya Institute of Technology |
Complex-valued Multilayer Perceptron Search Unilizing Eigen Vector Descent and Reducibility Mapping Shinya Suzumura, Ryohei Nakano (Chubu Univ.) NC2011-88 |
A complex-valued multilayer perceptron (MLP) can approximate a periodic or unbounded function, which cannot be easily re... [more] |
NC2011-88 pp.25-30 |
NC, MBE [detail] |
2010-12-19 11:20 |
Aichi |
Nagoya Univ. |
Search Method Utilizing Singular Region of Multilayer Perceptron Seiya Satoh, Takayuki Ohwaki, Ryohei Nakano (Chubu Univ.) MBE2010-70 NC2010-81 |
In a search space of MLP(J), multi-layer perceptron having J hidden units, there exists a singular region created by the... [more] |
MBE2010-70 NC2010-81 pp.85-90 |
NC, MBE (Joint) |
2008-12-20 10:00 |
Aichi |
Nagoya Inst. Tech. |
Clustering complex networks with the prior based on degree distribution Naoyuki Harada, Ichiro Takeuchi (NIT), Ryohei Nakano (Chubu Univ.) NC2008-73 |
Newman et al. proposed a graph clustering method based on a robabilistic mixture model with only the general assumption ... [more] |
NC2008-73 pp.1-6 |
NC, MBE (Joint) |
2008-12-20 11:05 |
Aichi |
Nagoya Inst. Tech. |
A Study on Variational Bayes Method with the Primitive Initial Point Yuta Ishikawa, Ichiro Takeuchi (NIT), Ryohei Nakano (Chubu Univ.) NC2008-75 |
The variational bayes (VB) method is widely used as an approximation of
the bayes method.
Since the objective functio... [more] |
NC2008-75 pp.13-18 |
NC, MBE (Joint) |
2008-12-20 14:30 |
Aichi |
Nagoya Inst. Tech. |
Gradient Based Two Dimensional Path Following for Kernel Machines Masayuki Karasuyama, Ichiro Takeuchi (NIT), Ryohei Nakano (Chubu Univ.) NC2008-80 |
The performance of the Kernel Machines depends on its hyperparameters such as a regularization parameter.
Since the pro... [more] |
NC2008-80 pp.43-48 |
MBE, NC (Joint) |
2007-12-22 09:25 |
Aichi |
|
Obtaining EM Initial Points by Using the Primitive Initial Point and Subsampling Strategy Yuta Ishikawa, Ryohei Nakano (Nagoya Inst. of Tech.) NC2007-72 |
The EM algorithm is an efficient algorithm to obtain the ML estimate for incomplete data, but has the local optimality p... [more] |
NC2007-72 pp.7-12 |
MBE, NC (Joint) |
2007-12-22 09:50 |
Aichi |
|
Optimizing SVR Hyperparameters via Fast Cross-Validation Masayuki Karasuyama, Ryohei Nakano (Nagoya Inst. of Tech.) NC2007-73 |
The performance of Support Vector Regression (SVR) deeply depends on its hyperparameters such as an insensitive zone thi... [more] |
NC2007-73 pp.13-18 |
NC |
2007-01-25 17:30 |
Hokkaido |
Noboribetsu Manseikaku(Noboribetsu) |
Adaptation of a Reinforcement Learning System IPMBN Using a Clustering Algorithm to Environmental Changes Daisuke Kitakoshi (Nagoya Inst. of Tech.), Terumasa Yamaguchi, Hiroyuki Shioya (Muroran Inst. of Tech.), Ryohei Nakano (Nagoya Inst. of Tech.) |
[more] |
NC2006-99 pp.65-70 |
NC |
2007-01-26 15:00 |
Hokkaido |
Noboribetsu Manseikaku(Noboribetsu) |
A Method for Simplifying Network Structure to Improve Efficiency in the Loopy-BP Algorithm Shunsuke Minamikawa, Daisuke Kitakoshi, Ryohei Nakano (Nagoya Inst. of Tech.) |
[more] |
NC2006-112 pp.69-74 |
NC, MBE (Joint) |
2006-12-05 14:50 |
Aichi |
Toyohashi Univ. of Tech. |
Multiple Regression with Automatic Nominal Space Partition using a Four-Layer Perceptron Yusuke Tanahashi, Yan Ying, Ryohei Nakano (Nagoya Inst. of Tech.) |
[more] |
NC2006-82 pp.67-72 |
NC, MBE (Joint) |
2006-12-05 15:10 |
Aichi |
Toyohashi Univ. of Tech. |
Competition for Survival of Rules Representing Correct Niches Takayuki Semba, Daisuke Kitakoshi, Ryohei Nakano (Nagoya Inst.) |
[more] |
NC2006-83 pp.73-78 |