CS 723 Reading List

  • Please submit your critiques before noon on Mondays to cs723-grading@groupes.epfl.ch
  • Week
    #1Sep. 18: Introduction
    PDF The Task of the Referee
    #2Sep. 25: Benchmarks and Analytics
    PDF DAWNBench: An End-to-End Deep Learning Benchmark and Competition [Stanford, NIPS 17]
    PDF CLARINET: WAN-Aware Optimization for Analytics Queries [UW-Madison, OSDI 16]
    #3Oct. 02: Parameter Servers
    PDF Scaling Distributed Machine Learning with the Parameter Server [CMU, OSDI 14]
    PDF GeePS: Scalable deep learning on distributed GPUs with a GPU-specialized parameter server [CMU, EuroSys 16]
    #4Oct. 09: Large Scale ML Frameworks
    PDF Project Adam: Building an Efficient and Scalable Deep Learning Training System [Microsoft, OSDI 14]
    PDF TensorFlow: A System for Large-Scale Machine Learning [Google, OSDI 16]
    #5Oct. 16: Distributed Training with GPUs and ML for systems
    PDF The Case for Learned Index Structures
    PDF FireCaffe: near-linear acceleration of deep neural network training on compute clusters [DeepScale, CVPR 16]
    #6Oct. 23: Distributed Stochastic Gradient Descent
    PDF Parle: parallelizing stochastic gradient descent [UCLA, SysML 18]
    PDF Decoupled Parallel Backpropagation with Convergence Guarantee [University of Pittsburgh, ICML 18]
    #7Oct. 29: Training with Low-precision Gradients
    PDF SIGNSGD: Compressed Optimisation for Non-Convex Problems [Caltech, ICML 18]
    PDF Deep Gradient Compression: Reducing the Communication Bandwidth for Distributed Training [Tsinghua, ICLR 18]
    #8Nov. 05: Training with Low-precision Gradients
    PDF QSGD: Communication-Efficient SGD via Gradient Quantization and Encoding [IST Austria & ETH, NIPS 17]
    PDF Error Compensated Quantized SGD and its Applications to Large-scale Distributed Optimization [Tencent, ICML 18]
    #9Nov. 12: Deep Learning with Low-precision Computations
    PDF Flexpoint: An Adaptive Numerical Format for Efficient Training of Deep Neural Networks [Intel, NIPS 17]
    PDF EIE: Efficient Inference Engine on Compressed Deep Neural Network [Stanford, ISCA 16]
    #10Nov. 19: Distributed IMAGENET Training
    ImageNet Training in 18 Minutes - https://www.fast.ai/2018/08/10/fastai-diu-imagenet/
    PDF Speeding up ImageNet Training on Supercomputers [UCB, SysML 18]
    #11Nov. 26: Domain Specific Languages for ML
    PDF Scale-out acceleration for machine learning [Georgia Tech, MICRO 17]
    PDF TVM: An Automated End-to-End Optimizing Compiler for Deep Learning [U. of Washington, SysML 18]
    #12Dec. 03: Domain Specific Languages for ML
    PDF Dynamic Control Flow in Large-Scale Machine Learning [Microsoft, EuroSys 18]
    PDF RLlib: Abstractions for Distributed Reinforcement Learning [UCB, ICML 18]
    #13Dec. 10: Hardware Accelerators for Deep Learning
    PDF In-Datacenter Performance Analysis of a Tensor Processing Unit [Google, ISCA 17]
    PDF Eyeriss: An Energy-Efficient Reconfigurable Accelerator for Deep Convolutional Neural Networks [MIT, IEEE Journal of Solid-State Circuits]
    #14Dec. 17: Security
    PDF DeepXplore: Automated Whitebox Testing of Deep Learning Systems [Columbia University, SOSP 17]
    PDF Stealing Machine Learning Models via Prediction APIs [EPFL, USENIX Security, 2016]