{"author_url":"https://blog.hatena.ne.jp/kitanokumo/","width":"100%","published":"2016-12-05 12:58:03","height":"190","provider_url":"https://hatena.blog","categories":["chainer"],"url":"https://studylog.hateblo.jp/entry/2016/12/05/125803","title":"\u4eca\u5e74(2016)\u53c2\u8003\u306b\u3055\u305b\u3066\u3082\u3089\u3063\u305fChainer\u306e\u8ad6\u6587\u5b9f\u88c5\u3001\u30b5\u30f3\u30d7\u30eb\u30b3\u30fc\u30c9\u96c6","type":"rich","blog_url":"https://studylog.hateblo.jp/","image_url":"https://cdn-ak.f.st-hatena.com/images/fotolife/k/kitanokumo/20161201/20161201131745.png","blog_title":"studylog/\u5317\u306e\u96f2","version":"1.0","description":"\u53bb\u5e74\u66f8\u3044\u305f\u30b5\u30f3\u30d7\u30eb\u30b3\u30fc\u30c9\u96c6\u306e2016\u5e74\u7248\u3067\u3059\u3002 \u500b\u4eba\u7684\u306a\u8208\u5473\u7bc4\u56f2\u306e\u307f\u96c6\u3081\u3066\u3044\u308b\u306e\u3067\u7db2\u7f85\u7684\u3067\u306f\u7121\u3044\u3068\u306f\u601d\u3044\u307e\u3059\u3002 \u57fa\u672c\u7684\u306b\u4e0a\u306e\u65b9\u304c\u65b0\u3057\u3044\u30b3\u30fc\u30c9\u3067\u3059\u3002 QRNN(Quasi-Recurrent Neural Networks) \u8ad6\u6587\u3067\u306fchainer\u3092\u4f7f\u3063\u3066\u5b9f\u9a13\u3057\u3066\u304a\u308a\u3001\u666e\u901a\u306eLSTM\u306f\u3082\u3061\u308d\u3093cuDNN\u3092\u4f7f\u3063\u305fLSTM\u3088\u308a\u3082\u9ad8\u901f\u3089\u3057\u3044\u3002 \u4e00\u756a\u4e0b\u306bchainer\u5b9f\u88c5\u30b3\u30fc\u30c9\u304c\u57cb\u3081\u8fbc\u307e\u308c\u3066\u3044\u308b\u3002 New neural network building block allows faster and more accurate text understanding[1611.01576] Quasi-R\u2026","author_name":"kitanokumo","provider_name":"Hatena Blog","html":""}