ICML 2018 深度学习论文及代码集锦（1）

[1]Spline Filters For End-to-End Deep Learning

Randall Balestriero et al.

Rice University

http://proceedings.mlr.press/v80/balestriero18a/balestriero18a.pdf

这篇论文的贡献在于

各方法效果对比如下

代码地址

https://github.com/RandallBalestriero/SplineWavelet

[2]Predict and Constrain: Modeling Cardinality in Deep Structured Prediction

Nataly Brukhim

Tel Aviv University

http://proceedings.mlr.press/v80/brukhim18a/brukhim18a.pdf

模型结构示例如下

映射算法伪代码如下

各算法效果对比如下

其中上表中的部分算法解释如下

代码地址

https://github.com/Natalybr/predict_and_constrain

[3]Learning Deep ResNet Blocks Sequentially using Boosting Theory

Furong Huang et al.

University of Maryland, Princeton University, Microsoft Research.

http://proceedings.mlr.press/v80/huang18b/huang18b.pdf

ResNet网络结构如下

BoostResNet的算法伪代码如下

算法效果对比如下

[4]Self-Consistent Trajectory Autoencoder: Hierarchical Reinforcement Learningwith Trajectory Embeddings

John D. Co-Reyes et al.

University of California, Berkeley; Google Brain

http://proceedings.mlr.press/v80/co-reyes18a/co-reyes18a.pdf

状态及策略解码图示如下

SeCTAR（self-consistent trajectory autoencoder）图示如下

算法伪代码如下

代码地址

https://github.com/wyndwarrior/Sectar

[5]Deep Reinforcement Learning in Continuous Action Spaces:a Case Study in the Game of Simulated Curling

Kyowoon Lee et al.

Ulsan National Institute of Science and Technology, Korea University

http://proceedings.mlr.press/v80/lee18b/lee18b.pdf

网络结构如下

算法伪代码如下

代码地址

https://github.com/leekwoon/KR-DL-UCT

[6]Dropout Training, Data-dependent Regularization, and Generalization Bounds

Wenlong Mou et al.

University of California, Berkeley; University of Wisconsin, Madison; PekingUniversity

http://proceedings.mlr.press/v80/mou18a/mou18a.pdf

算法伪代码如下

各方法效果对比如下

您可能感兴趣