Skip to content

Latest commit

 

History

History
45 lines (36 loc) · 2.13 KB

README.md

File metadata and controls

45 lines (36 loc) · 2.13 KB

Intro

I am Jiecao YU, a PhD student in the University of Michigan.

This is a paper reading list about DNN acceleration (also general Deep Learning).

Still under construction.

DNN Acceleration

DNN architecture

  • MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications [ paper | 20170612001 ]
  • LightRNN: Memory and Computation-Efficient Recurrent Neural Networks [ paper | 20170612005 ]
  • ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices [ paper | 20170713001 ]

Accelerators

  • Minerva: Enabling Low-Power, Highly-Accurate Deep Neural Network Accelerators [ paper | 20170630001 ]

DNN pruning

  • Pruning Convolutional Neural Networks for Resource Efficient Transfer Learning [ paper | 20170612002 ]
  • Learning Structured Sparsity in Deep Neural Networks [ paper | 20170612003 ]
  • Dynamic Network Surgery for Efficient DNNs [ paper | 20170612004 ]
  • Pruning Filters for Efficient ConvNets [ paper | 20170901001 ]
  • Data-Driven Sparse Structure Selection for Deep Neural Networks [ paper | 20170901002 ]
  • ThiNet: A Filter Level Pruning Method for Deep Neural Network Compression [ paper | 20170901003 ]
  • Scalpel: Customizing DNN Pruning to the Underlying Hardware Parallelism [ paper | 20170901004 ]

Binarized neural network

  • Binarized Neural Networks: Training Deep Neural Networks with Weights and Activations Constrained to +1 or -1 [ paper | 20170901005 ]
  • XNOR-Net: ImageNet Classification Using Binary Convolutional Neural Networks [ paper | 20170901006 ]

Others