Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox
Picture for Shuang Song

Shuang Song

Practical and Private (Deep) Learning without Sampling or Shuffling


Feb 26, 2021
Peter Kairouz, Brendan McMahan, Shuang Song, Om Thakkar, Abhradeep Thakurta, Zheng Xu

Figure 1 for Practical and Private (Deep) Learning without Sampling or Shuffling
Figure 2 for Practical and Private (Deep) Learning without Sampling or Shuffling
Figure 3 for Practical and Private (Deep) Learning without Sampling or Shuffling
Figure 4 for Practical and Private (Deep) Learning without Sampling or Shuffling

  Access Paper or Ask Questions

Adversary Instantiation: Lower Bounds for Differentially Private Machine Learning


Jan 11, 2021
Milad Nasr, Shuang Song, Abhradeep Thakurta, Nicolas Papernot, Nicholas Carlini

Figure 1 for Adversary Instantiation: Lower Bounds for Differentially Private Machine Learning
Figure 2 for Adversary Instantiation: Lower Bounds for Differentially Private Machine Learning
Figure 3 for Adversary Instantiation: Lower Bounds for Differentially Private Machine Learning
Figure 4 for Adversary Instantiation: Lower Bounds for Differentially Private Machine Learning

  Access Paper or Ask Questions

An Attack on InstaHide: Is Private Learning Possible with Instance Encoding?


Nov 10, 2020
Nicholas Carlini, Samuel Deng, Sanjam Garg, Somesh Jha, Saeed Mahloujifar, Mohammad Mahmoody, Shuang Song, Abhradeep Thakurta, Florian Tramer

Figure 1 for An Attack on InstaHide: Is Private Learning Possible with Instance Encoding?
Figure 2 for An Attack on InstaHide: Is Private Learning Possible with Instance Encoding?
Figure 3 for An Attack on InstaHide: Is Private Learning Possible with Instance Encoding?
Figure 4 for An Attack on InstaHide: Is Private Learning Possible with Instance Encoding?

  Access Paper or Ask Questions

Tempered Sigmoid Activations for Deep Learning with Differential Privacy


Jul 28, 2020
Nicolas Papernot, Abhradeep Thakurta, Shuang Song, Steve Chien, Úlfar Erlingsson

Figure 1 for Tempered Sigmoid Activations for Deep Learning with Differential Privacy
Figure 2 for Tempered Sigmoid Activations for Deep Learning with Differential Privacy
Figure 3 for Tempered Sigmoid Activations for Deep Learning with Differential Privacy
Figure 4 for Tempered Sigmoid Activations for Deep Learning with Differential Privacy

  Access Paper or Ask Questions

Characterizing Private Clipped Gradient Descent on Convex Generalized Linear Problems


Jun 11, 2020
Shuang Song, Om Thakkar, Abhradeep Thakurta


  Access Paper or Ask Questions

Combining MixMatch and Active Learning for Better Accuracy with Fewer Labels


Dec 03, 2019
Shuang Song, David Berthelot, Afshin Rostamizadeh

Figure 1 for Combining MixMatch and Active Learning for Better Accuracy with Fewer Labels
Figure 2 for Combining MixMatch and Active Learning for Better Accuracy with Fewer Labels
Figure 3 for Combining MixMatch and Active Learning for Better Accuracy with Fewer Labels
Figure 4 for Combining MixMatch and Active Learning for Better Accuracy with Fewer Labels

  Access Paper or Ask Questions

That which we call private


Aug 08, 2019
Úlfar Erlingsson, Ilya Mironov, Ananth Raghunathan, Shuang Song


  Access Paper or Ask Questions

A Comparison of Stereo-Matching Cost between Convolutional Neural Network and Census for Satellite Images


May 22, 2019
Bihe Chen, Rongjun Qin, Xu Huang, Shuang Song, Xiaohu Lu

Figure 1 for A Comparison of Stereo-Matching Cost between Convolutional Neural Network and Census for Satellite Images
Figure 2 for A Comparison of Stereo-Matching Cost between Convolutional Neural Network and Census for Satellite Images
Figure 3 for A Comparison of Stereo-Matching Cost between Convolutional Neural Network and Census for Satellite Images
Figure 4 for A Comparison of Stereo-Matching Cost between Convolutional Neural Network and Census for Satellite Images

  Access Paper or Ask Questions

Scalable Private Learning with PATE


Feb 24, 2018
Nicolas Papernot, Shuang Song, Ilya Mironov, Ananth Raghunathan, Kunal Talwar, Úlfar Erlingsson

Figure 1 for Scalable Private Learning with PATE
Figure 2 for Scalable Private Learning with PATE
Figure 3 for Scalable Private Learning with PATE
Figure 4 for Scalable Private Learning with PATE
* Published as a conference paper at ICLR 2018 
  Access Paper or Ask Questions

Rényi Differential Privacy Mechanisms for Posterior Sampling


Oct 02, 2017
Joseph Geumlek, Shuang Song, Kamalika Chaudhuri

Figure 1 for Rényi Differential Privacy Mechanisms for Posterior Sampling
Figure 2 for Rényi Differential Privacy Mechanisms for Posterior Sampling
Figure 3 for Rényi Differential Privacy Mechanisms for Posterior Sampling
* to be published in NIPS 2017 
  Access Paper or Ask Questions

Composition Properties of Inferential Privacy for Time-Series Data


Jul 10, 2017
Shuang Song, Kamalika Chaudhuri


  Access Paper or Ask Questions

Pufferfish Privacy Mechanisms for Correlated Data


Mar 12, 2017
Shuang Song, Yizhen Wang, Kamalika Chaudhuri

Figure 1 for Pufferfish Privacy Mechanisms for Correlated Data
Figure 2 for Pufferfish Privacy Mechanisms for Correlated Data
Figure 3 for Pufferfish Privacy Mechanisms for Correlated Data
Figure 4 for Pufferfish Privacy Mechanisms for Correlated Data

  Access Paper or Ask Questions

Learning from Data with Heterogeneous Noise using SGD


Dec 17, 2014
Shuang Song, Kamalika Chaudhuri, Anand D. Sarwate

Figure 1 for Learning from Data with Heterogeneous Noise using SGD

  Access Paper or Ask Questions

The Large Margin Mechanism for Differentially Private Maximization


Sep 07, 2014
Kamalika Chaudhuri, Daniel Hsu, Shuang Song


  Access Paper or Ask Questions