Privacy Risks of Securing Machine Learning Models against Adversarial Examples

24 May 20191 code implementation

To perform the membership inference attacks, we leverage the existing inference methods that exploit model predictions.

ADVERSARIAL DEFENSE INFERENCE ATTACK

Information-Theoretic Understanding of Population Risk Improvement with Model Compression

27 Jan 20191 code implementation

We show that model compression can improve the population risk of a pre-trained model, by studying the tradeoff between the decrease in the generalization error and the increase in the empirical risk with model compression.

MODEL COMPRESSION REGRESSION

Distill-and-Compare: Auditing Black-Box Models Using Transparent Model Distillation

17 Oct 20173 code implementations

We compare the student model trained with distillation to a second un-distilled transparent model trained on ground-truth outcomes, and use differences between the two models to gain insight into the black-box model.

Random Erasing Data Augmentation

16 Aug 20176 code implementations

In this paper, we introduce Random Erasing, a new data augmentation method for training the convolutional neural network (CNN).

IMAGE AUGMENTATION IMAGE CLASSIFICATION OBJECT DETECTION PERSON RE-IDENTIFICATION

A Distributional Perspective on Reinforcement Learning

ICML 2017 10 code implementations

We obtain both state-of-the-art results and anecdotal evidence demonstrating the importance of the value distribution in approximate reinforcement learning.

ATARI GAMES

Orthogonal Statistical Learning

25 Jan 20191 code implementation

We provide excess risk guarantees for statistical learning in a setting where the population risk with respect to which we evaluate the target model depends on an unknown model that must be to be estimated from data (a "nuisance model").

DOMAIN ADAPTATION

ECO: Efficient Convolution Operators for Tracking

CVPR 2017 2 code implementations

Moreover, our fast variant, using hand-crafted features, operates at 60 Hz on a single CPU, while obtaining 65. 0% AUC on OTB-2015.

VISUAL OBJECT TRACKING

Unrestricted Adversarial Examples

22 Sep 20181 code implementation

We introduce a two-player contest for evaluating the safety and robustness of machine learning systems, with a large prize pool.

Parallelizing Stochastic Gradient Descent for Least Squares Regression: mini-batching, averaging, and model misspecification

12 Oct 20161 code implementation

In particular, this work provides a sharp analysis of: (1) mini-batching, a method of averaging many samples of a stochastic gradient to both reduce the variance of the stochastic gradient estimate and for parallelizing SGD and (2) tail-averaging, a method involving averaging the final few iterates of SGD to decrease the variance in SGD's final iterate.

REGRESSION