講演抄録/キーワード |
講演名 |
2014-11-17 17:00
[ポスター講演]Regularized multi-task learning for multi-dimensional log-density gradient estimation ○Ikko Yamane(Tokyo Inst. of Tech.)・Hiroaki Sasaki・Masashi Sugiyama(Univ. of Tokyo) IBISML2014-58 |
抄録 |
(和) |
(事前公開アブストラクト) Least-squares log-density gradient (LSLDG) is a recently proposed method to directly estimate the gradients of log-densities. In this work, we apply regularized multi-task learning to multi-dimensional LSLDG and show its usefulness. |
(英) |
Log-density gradient estimation is a fundamental statistical problem and it has various practical applications such as clustering and a measure for non-Gaussianity. A naive two-step approach of rst estimating
the density and then taking its log-gradient does not perform well because an accurate density estimate does not
necessarily lead to an accurate log-density gradient estimate. To cope with this problem, a method to directly estimate the log-density gradient without density estimation was explored. However, even with the direct estimator,
high-dimensional log-density gradient estimation is still challenging. In this paper, we propose to apply regularized
multi-task learning to direct log-density gradient estimation and show its usefulness experimentally. |
キーワード |
(和) |
/ / / / / / / |
(英) |
multi-task learning / log-density gradient estimation / / / / / / |
文献情報 |
信学技報, vol. 114, no. 306, IBISML2014-58, pp. 177-183, 2014年11月. |
資料番号 |
IBISML2014-58 |
発行日 |
2014-11-10 (IBISML) |
ISSN |
Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 |
著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します.(許諾番号:10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
PDFダウンロード |
IBISML2014-58 |