arXiv research

Generative Intervention Models predict perturbation effects without knowing the underlying mechanisms.

The paper tackles extrapolation of gene knockouts effects on RNA counts.

Study reveals class-dependent effects in perturbation-based feature attribution metrics for time series classification.

New method for analyzing compositional data, addressing biases in summary statistics.

New method $μP^2$ improves neural network training by scaling perturbations layerwise.

SAMS-VAE models cellular perturbations using sparse additive mechanisms.

DBPA assesses LLM perturbations using frequentist hypothesis testing.

Perturbative string amplitudes are correctly derived from the string geometry theory, which is one of the candidates of a non-perturbative formulation of string theory. In order to derive non-perturbative effects rather easily, we formulate topological string geometry theory. We derive the perturbative partition functi…

The problem of attribution is concerned with identifying the parts of an input that are responsible for a model's output. An important family of attribution methods is based on measuring the effect of perturbations applied to the input. In this paper, we discuss some of the shortcomings of existing approaches to pertur…

Wide networks learn from adversarial perturbations effectively.

TULiP estimates uncertainty for deep learning models safely.

This paper presents a new approach, called perturb-max, for high-dimensional statistical inference that is based on applying random perturbations followed by optimization. This framework injects randomness to maximum a-posteriori (MAP) predictors by randomly perturbing the potential function for the input. A classic re…

Classifiers such as deep neural networks have been shown to be vulnerable against adversarial perturbations on problems with high-dimensional input space. While adversarial training improves the robustness of image classifiers against such adversarial perturbations, it leaves them sensitive to perturbations on a non-ne…

DEceit constructs effective universal pixel-restricted perturbations for deep image classifiers.

Paper proposes a method to generate adversarial perturbations for black-box attacks without accessing inner states.

New method defines GCM spheres in Kerr perturbations, proving their stability.

Adversarial attacks for image classification are small perturbations to images that are designed to cause misclassification by a model. Adversarial attacks formally correspond to an optimization problem: find a minimum norm image perturbation, constrained to cause misclassification. A number of effective attacks have b…

One of the most pressing questions in climate science is that of the effect of anthropogenic aerosol on the Earth's energy balance. Aerosols provide the `seeds' on which cloud droplets form, and changes in the amount of aerosol available to a cloud can change its brightness and other physical properties such as optical…

CAT improves robustness of neural networks by customizing perturbation levels.

Minimal token perturbations reveal how Transformer models process information.

Improved online Lasso reduces regret in sparse linear contextual bandits.

Study small perturbations on low energy Laplace eigenfunctions.

While several feature scoring methods are proposed to explain the output of complex machine learning models, most of them lack formal mathematical definitions. In this study, we propose a novel definition of the feature score using the maximally invariant data perturbation, which is inspired from the idea of adversaria…

CausalRegNet generates accurate data for gene perturbation experiments, improving CSL methods.

Carefully crafted, often imperceptible, adversarial perturbations have been shown to cause state-of-the-art models to yield extremely inaccurate outputs, rendering them unsuitable for safety-critical application domains. In addition, recent work has shown that constraining the attack space to a low frequency regime is …

Framework for causal discovery using multi-modal data.

Deep neural networks (DNNs) have achieved superior performance in various prediction tasks, but can be very vulnerable to adversarial examples or perturbations. Therefore, it is crucial to measure the sensitivity of DNNs to various forms of perturbations in real applications. We introduce a novel perturbation manifold …

This work introduces adversarial sparsity to measure robustness beyond adversarial accuracy.

We consider problems in which a system receives external \emph{perturbations} from time to time. For instance, the system can be a train network in which particular lines are repeatedly disrupted without warning, having an effect on passenger behavior. The goal is to predict changes in the behavior of the system at par…

Stochastic neural net weights are used in a variety of contexts, including regularization, Bayesian neural nets, exploration in reinforcement learning, and evolution strategies. Unfortunately, due to the large number of weights, all the examples in a mini-batch typically share the same weight perturbation, thereby limi…

ALPS improves neural network robustness and generalization.

Develops new methods to create imperceptible image changes that fool classifiers.

Adversarial attacks add perturbations to the input features with the intent of changing the classification produced by a machine learning system. Small perturbations can yield adversarial examples which are misclassified despite being virtually indistinguishable from the unperturbed input. Classifiers trained with stan…

This paper proposes a new algorithm for controlling classification results by generating a small additive perturbation without changing the classifier network. Our work is inspired by existing works generating adversarial perturbation that worsens classification performance. In contrast to the existing methods, our wor…

Simple regional perturbations maintain model transferability while reducing adversarial example distortion.

The paper learns perturbation sets from data to improve robustness in machine learning.

Adversarial fog tests autonomous navigation models.

New method estimates heterogeneous treatment effects with improved guarantees.

We investigate the generalizability of deep learning based on the sensitivity to input perturbation. We hypothesize that the high sensitivity to the perturbation of data degrades the performance on it. To reduce the sensitivity to perturbation, we propose a simple and effective regularization method, referred to as spe…

The paper assesses machine learning robustness with covariate perturbations.

Paper analyzes GCNN sensitivity to probabilistic graph perturbations.

Improved generalization with semantic perturbations using normalizing flows.

Study privacy and accuracy in high-dimensional LASSO with perturbation mechanisms.

Beyond existing multi-view clustering, this paper studies a more realistic clustering scenario, referred to as incomplete multi-view clustering, where a number of data instances are missing in certain views. To tackle this problem, we explore spectral perturbation theory. In this work, we show a strong link between per…

Despite the tremendous success of deep neural networks in various learning problems, it has been observed that adding an intentionally designed adversarial perturbation to inputs of these architectures leads to erroneous classification with high confidence in the prediction. In this work, we propose a general framework…

PerturBench benchmarks ML models for cellular perturbation analysis.

We propose denoising dictionary learning (DDL), a simple yet effective technique as a protection measure against adversarial perturbations. We examined denoising dictionary learning on MNIST and CIFAR10 perturbed under two different perturbation techniques, fast gradient sign (FGSM) and jacobian saliency maps (JSMA). W…

PEP improves deep network performance and calibration by perturbing optimal parameters.