イベントJOIN US

第12回データサイエンスセミナー(10/19, 参加自由) 15:30開始に変更 Aapo Hyvärinen 氏

2016年10月07日

日時:10月19日(水) 15:30~16:30 + ε (時間が変更になりました)
場所:DSラーニングコモンズ(データサイエンス棟1F)
アクセス: https://www.ds.shiga-u.ac.jp/access/

講演者:Aapo Hyvärinen 氏 (Gatsby Computational Neuroscience Unit, University College London / Department of Computer Science, University of Helsinki)

題目:「Unsupervised Feature Extraction by Time-Contrastive Learning and Nonlinear ICA」

概要:Nonlinear independent component analysis (ICA) provides an appealing framework for unsupervised feature learning, but the models proposed so far are not identifiable. Here, we first propose a new intuitive principle of unsupervised deep learning from time series which uses the nonstationary structure of the data. Our learning principle, time-contrastive learning (TCL), finds a representation which allows optimal discrimination of time segments (windows). Surprisingly, we show how TCL can be related to a nonlinear ICA model, when ICA is redefined to include temporal nonstationarities. In particular, we show that TCL combined with linear ICA estimates the nonlinear ICA model up to point-wise transformations of the sources, and this solution is unique — thus providing the first identifiability result for nonlinear ICA which is rigorous, constructive, as well as very general. https://arxiv.org/pdf/1605.06336v1.pdf
---
滋賀大学データサイエンス教育研究センターのページはコチラ