[Tue, 1 Dec 2015 ~ Fri, 4 Dec 2015] Deep Learning in arxiv

AnalyzingClassifiers: Fisher Vectors and Deep Neural Networks

fv与dnn特征进行比较，得出的结论是fv比较偏向上下文信息，而dnn只比较关注物体本身信息。

 

TowardsDropout Training for Convolutional Neural Networks

除了dropout神经元之外，是否还可以dropoutpool, fully connection, convolution ?

该文章做了一些对比。

这个是比较好的一个工程点。

 

 

 

Rethinkingthe Inception Architecture for Computer Vision

 

 

5x5conv->3x3 conv+3x3 fc :we end up with a net 18 25× reduction of computation,resulting in a relative gain of 28% by this factorization

 

Instead,we argue that the auxiliary classifiers act as regularizer

Thisis supported by the fact that the main classifier of the network performsbetter if the side branch is batch-normalized [7] or has a dropout layer. Thisalso gives a weak supporting evidence for the conjecture that batchnormalization acts as a regularizer

 

Inaddition, gradient clipping [14] was found to be useful to stabilize thetraining

 

Googlenet特点： Much of the original gains of theGoogLeNet network [20] arise from a very generous use of dimension reduction.

 

网络结构：

结果对比：

 

 

MXNet:A Flexible and Efficient Machine Learning Library for Heterogeneous DistributedSystems

深度学习并行化训练相关代码

 

 

 

 

 

Attribute2Image:Conditional Image Generation from Visual Attributes

比较有意思的限定场景下的，attribute2image实验以及应用方向

1. search

2. image generation

 

Actions∼Transformations

Inthis paper, we propose a novel representation for actions by modeling action asa transformation which changes the state of the environment before the actionhappens (precondition) to the state after the action (effect)

蛮优雅的一个网络结构