arXiv research

New framework for understanding adversarial and stochastic learning.

We investigate the non-identifiability issues associated with bidirectional adversarial training for joint distribution matching. Within a framework of conditional entropy, we propose both adversarial and non-adversarial approaches to learn desirable matched joint distributions for unsupervised and supervised tasks. We…

Adversarial training (AT) is among the most effective techniques to improve model robustness by augmenting training data with adversarial examples. However, most existing AT methods adopt a specific attack to craft adversarial examples, leading to the unreliable robustness against other unseen attacks. Besides, a singl…

RATIO improves neural network robustness and explainability.

Modern machine learning algorithms perform poorly on adversarially manipulated data. Adversarial risk quantifies the error of classifiers in adversarial settings; adversarial classifiers minimize adversarial risk. In this paper, we analyze adversarial risk and adversarial classifiers from an optimal transport perspecti…

Proposes a non-adversarial method for distribution matching.

Proposes BATer for improved adversarial example detection.

The recent advances in deep transfer learning reveal that adversarial learning can be embedded into deep networks to learn more transferable features to reduce the distribution discrepancy between two domains. Existing adversarial domain adaptation methods either learn a single domain discriminator to align the global …

New adversarial attack method based on deep feature distributions.

Neural networks are vulnerable to small adversarial perturbations. Existing literature largely focused on understanding and mitigating the vulnerability of learned models. In this paper, we demonstrate an intriguing phenomenon about the most popular robust training method in the literature, adversarial training: Advers…

XEnsemble improves DNN robustness against adversarial and out-of-distribution inputs.

TAET tackles long-tailed distributions in adversarial robustness.

Adversarial training achieves optimal test error for shallow networks.

This paper explores tradeoffs between standard and adversarial risks in distributionally adversarial training.

We present a principled approach for detecting out-of-distribution (OOD) and adversarial samples in deep neural networks. Our approach consists in modeling the outputs of the various layers (deep features) with parametric probability distributions once training is completed. At inference, the likelihoods of the deep fe…

Recent work on adversarial attack has shown that Projected Gradient Descent (PGD) Adversary is a universal first-order adversary, and the classifier adversarially trained by PGD is robust against a wide range of first-order attacks. It is worth noting that the original objective of an attack/defense model relies on a d…

KSGAN uses KS distance for deep generative modeling.

Combines adversarial and interventional robustness for machine learning models.

The paper addresses adversarial robustness in in-context learning models.

New algorithm identifies near-optimal policies in adversarial distributed RL settings.

Algorithm learns without knowing distribution, reducing error.

A large body of recent work has investigated the phenomenon of evasion attacks using adversarial examples for deep learning systems, where the addition of norm-bounded perturbations to the test inputs leads to incorrect output classification. Previous work has investigated this phenomenon in closed-world systems where …

SGD-trained neural networks generalize well even with adversarial label noise.

This paper extends adversarial attacks to produce desired class probability distributions.

RIA method improves OoD generalization for covariate shift.

This paper examines how adversarial perturbations affect model performance and equilibrium learning.

Study analyzes adversarial training dynamics without data distribution assumptions.

The paper improves smoothed analysis for online problems with adaptive adversaries.

Algorithm improves online learning in adversarial bandits.

Paper shows MoM is optimal under adversarial contamination for certain distributions.

Robustly estimates linear regression coefficients with adversarial and noisy data.

A new model for sequential prediction handles adversarial examples by allowing abstention.

ALFI improves likelihood-free inference for black-box generators.

Adversarial training improves robustness of halfspaces in noisy data.

SignSGD improves distributed learning by tolerating faulty devices, including Byzantine adversaries.

While the last few decades have witnessed a huge body of work devoted to inference and learning in distributed and decentralized setups, much of this work assumes a non-adversarial setting in which individual nodes---apart from occasional statistical failures---operate as intended within the algorithmic framework. In r…

The easiness at which adversarial instances can be generated in deep neural networks raises some fundamental questions on their functioning and concerns on their use in critical systems. In this paper, we draw a connection between over-generalization and adversaries: a possible cause of adversaries lies in models desig…

New insights into learning from distributional adversaries and private data.

Study improves distributed linear estimation under adversarial conditions.

Non-convex SGD learns halfspaces with adversarial label noise efficiently.

EIGAN learns private representations without centralized data, outperforming state-of-the-art.

Interpolating label noise makes models vulnerable to adversarial attacks.

We propose a robust adversarial prediction framework for general multiclass classification. Our method seeks predictive distributions that robustly optimize non-convex and non-continuous multiclass loss metrics against the worst-case conditional label distributions (the adversarial distributions) that (approximately) m…

Adversarial consistency depends on the uniqueness of adversarial Bayes classifiers.

Paper optimizes FTPL for adversarial and stochastic bandits with specific tail distributions.

We present a neural network architecture based upon the Autoencoder (AE) and Generative Adversarial Network (GAN) that promotes a convex latent distribution by training adversarially on latent space interpolations. By using an AE as both the generator and discriminator of a GAN, we pass a pixel-wise error function acro…

New method improves robustness of smoothed classifiers against adversarial attacks.

New findings show learnable distributions remain learnable even with noisy or adversarial perturbations.