arXiv research

The paper proposes a method to detect and filter noisy or mislabeled data using pointwise mutual information.

Study improves image classifier robustness to random p-norm corruptions.

Study robust mean estimation under coordinate-level corruptions using Hamming distance.

We introduce a new model of stochastic bandits with adversarial corruptions which aims to capture settings where most of the input follows a stochastic pattern but some fraction of it can be adversarially changed to trick the algorithm, e.g., click fraud, fake reviews and email spam. The goal of this model is to encour…

Over the last few years, the phenomenon of adversarial examples --- maliciously constructed inputs that fool trained machine learning models --- has captured the attention of the research community, especially when the adversary is restricted to small modifications of a correctly handled input. Less surprisingly, image…

We present a representation learning method that learns features at multiple different levels of scale. Working within the unsupervised framework of denoising autoencoders, we observe that when the input is heavily corrupted during training, the network tends to learn coarse-grained features, whereas when the input is …

Improved covariate shift handling with node-based Bayesian neural networks.

RGI improves robustness of GAN-inversion for image restoration and anomaly detection.

This paper aims to develop a new and robust approach to feature representation. Motivated by the success of Auto-Encoders, we first theoretical summarize the general properties of all algorithms that are based on traditional Auto-Encoders: 1) The reconstruction error of the input can not be lower than a lower bound, wh…

NoisyMix boosts model robustness to common corruptions.

Design of reliable systems must guarantee stability against input perturbations. In machine learning, such guarantee entails preventing overfitting and ensuring robustness of models against corruption of input data. In order to maximize stability, we analyze and develop a computationally efficient implementation of Jac…

Class selectivity affects robustness to corruptions but not to adversarial attacks.

Unsupervised learning is of growing interest because it unlocks the potential held in vast amounts of unlabelled data to learn useful representations for inference. Autoencoders, a form of generative model, may be trained by learning to reconstruct unlabelled input data from a latent representation space. More robust r…

Study robust learning of Lipschitz functions under corrupted binary signals.

Boosting framework improves adversarial robustness in deep learning.

Evolutionary algorithm finds optimal pixel perturbations to improve neural network generalization.

We give the first polynomial-time algorithm for performing linear or polynomial regression resilient to adversarial corruptions in both examples and labels. Given a sufficiently large (polynomial-size) training set drawn i.i.d. from distribution D and subsequently corrupted on some fraction of points, our algorithm out…

Self-supervision provides effective representations for downstream tasks without requiring labels. However, existing approaches lag behind fully supervised training and are often not thought beneficial beyond obviating or reducing the need for annotations. We find that self-supervision can benefit robustness in a varie…

This paper improves test-time adaptation for distribution shifts using confidence maximization and input transformation.

The paper addresses uncertainties in spectral clustering of corrupted data.

New approach improves model robustness and calibration in latent space.

New algorithm learns Bayesian networks robustly in nearly-linear time.

A robust loss for anomaly mitigation and unsupervised contamination classification

This paper considers the problem of matrix completion when some number of the columns are completely and arbitrarily corrupted, potentially by a malicious adversary. It is well-known that standard algorithms for matrix completion can return arbitrarily poor results, if even a single column is corrupted. One direct appl…

State-of-the art vision models can achieve superhuman performance on image classification tasks when testing and training data come from the same distribution. However, when models are tested on corrupted images (e.g. due to scale changes, translations, or shifts in brightness or contrast), performance degrades signifi…

Develops a comprehensive theory of corruption in supervised learning.

COLUMBUS discovers new features to improve domain generalization.

Simple AMP algorithm robust to adversarial corruption.

DRO-Augment framework enhances deep neural network robustness.

Robust method estimates state, input, and parameters of linear systems online.

New algorithm robust to label corruptions in active learning.

Study finds corruption negatively impacts firm performance.

Unified framework for corruption-robust linear bandits with optimal gap-dependent misspecification bounds.

Robust testing of sparse signals in corrupted data.

We report quantitative relations between corruption level and economic factors, such as country wealth and foreign investment per capita, which are characterized by a power law spanning multiple scales of wealth and investments per capita. These relations hold for diverse countries, and also remain stable over differen…

CUTS removes corruption from models without clean data, improving utility and security.

Sensor fusion has wide applications in many domains including health care and autonomous systems. While the advent of deep learning has enabled promising multi-modal fusion of high-level features and end-to-end sensor fusion solutions, existing deep learning based sensor fusion techniques including deep gating architec…

Study robust estimation under varying corruption probabilities in data.

Binary classification improves with a small fraction of corrupted labels.

RINS-T solves time series inverse problems robustly without pretraining.

Student-teacher learning improves generalization with noisy inputs.

We study the problem of corrupted sensing, a generalization of compressed sensing in which one aims to recover a signal from a collection of corrupted or unreliable measurements. While an arbitrary signal cannot be recovered in the face of arbitrary corruption, tractable recovery is possible when both signal and corrup…

New algorithm optimizes noisy, potentially corrupted functions.

PCA is a classical statistical technique whose simplicity and maturity has seen it find widespread use as an anomaly detection technique. However, it is limited in this regard by being sensitive to gross perturbations of the input, and by seeking a linear subspace that captures normal behaviour. The first issue has bee…

New algorithm reduces linear contextual bandit regret with adversarial corruption.

We extend the model of stochastic bandits with adversarial corruption (Lykouriset al., 2018) to the stochastic linear optimization problem (Dani et al., 2008). Our algorithm is agnostic to the amount of corruption chosen by the adaptive adversary. The regret of the algorithm only increases linearly in the amount of cor…

In this work, we initiate a formal study of probably approximately correct (PAC) learning under evasion attacks, where the adversary's goal is to \emph{misclassify} the adversarially perturbed sample point $\widetilde{x}$, i.e., $h(\widetilde{x})\neq c(\widetilde{x})$, where $c$ is the ground truth concept and $h$ is t…

Robust algorithm optimizes corrupted Gaussian process bandits.