随笔分类 - 深度学习

Lessons learned from manually classifying CIFAR-10

摘要：Lessons learned from manually classifying CIFAR-10Apr 27, 2011CIFAR-10Note, this post is from 2011 and slightly outdated in some places.Statistics. CI... 阅读全文

posted @ 2015-06-06 10:11 菜鸡一枚阅读(583) 评论(0) 推荐(0) 编辑

Why are Eight Bits Enough for Deep Neural Networks?

摘要：Why are Eight Bits Enough for Deep NeuralNetworks?Deep learning is a very weird technology. It evolved over decades on a very different track than the... 阅读全文

posted @ 2015-05-25 19:58 菜鸡一枚阅读(324) 评论(0) 推荐(0) 编辑

深度学习 vs. 概率图模型 vs. 逻辑学

摘要：深度学习 vs. 概率图模型 vs. 逻辑学摘要：本文回顾过去50年人工智能（AI）领域形成的三大范式：逻辑学、概率方法和深度学习。文章按时间顺序展开，先回顾逻辑学和概率图方法，然后就人工智能和机器学习的未来走向做些预测。【编者按】在上个月发表博客文章《深度学习 vs. 机器学习 vs. 模式识别》... 阅读全文

posted @ 2015-05-12 20:44 菜鸡一枚阅读(504) 评论(0) 推荐(0) 编辑

WHAT I READ FOR DEEP-LEARNING

摘要：WHAT I READ FOR DEEP-LEARNINGToday, I spent some time on two new papers proposing a new way of training very deep neural networks (Highway-Networks) a... 阅读全文

posted @ 2015-05-12 20:35 菜鸡一枚阅读(268) 评论(0) 推荐(0) 编辑

Initialization of deep networks

摘要：Initialization of deep networks24 Feb 2015Gustav LarssonAs we all know, the solution to a non-convex optimization algorithm (like stochastic gradient ... 阅读全文

posted @ 2015-05-02 13:09 菜鸡一枚阅读(525) 评论(0) 推荐(0) 编辑

DEEP LEARNING WITH STRUCTURE

摘要：DEEP LEARNING WITH STRUCTURECharlie Tang is aPhD studentin theMachine Learning groupat the University of Toronto, working withGeoffrey Hinton and Rusl... 阅读全文

posted @ 2015-05-02 12:25 菜鸡一枚阅读(448) 评论(0) 推荐(0) 编辑