Research
On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

169,341 papers · 148 categories

Trend · papers per month

81162242323 · Jun 202019922001200920182026
48 results for weight perturbation

AWP improves robustness by flattening weight loss landscape.

problem Improving robustness of deep neural networks against adversarial examples.
method Explicitly regularizes the flatness of weight loss landscape through adversarial weight perturbation.
result AWP forms a double-perturbation mechanism in adversarial training, leading to flatter weight loss landscape.

Adversarial weight perturbations can inject backdoors into trained neural models.

problem Security risk of using publicly available trained models due to backdoors.
method Extended adversarial perturbations to model weights, using a composite loss and projected gradient descent.
result Adversarial weight perturbations can be successfully injected with very small changes, exposing security risks across various tasks.

Flipout decorrelates mini-batch weights for more variance reduction.

problem Limited variance reduction in mini-batches due to shared weight perturbations.
method Implicitly samples pseudo-independent weight perturbations for each example.
result Achieves ideal linear variance reduction for various network types.

Efficiently learns perturb-and-map models using weighted log-likelihood.

problem Structured output prediction with weighted Hamming losses.
method Generalizes perturb-and-MAP framework, uses dynamic graph cuts for MAP inference, and double stochastic gradient descent for efficient learning.
result Shows efficiency in learning log-supermodular models with weak supervision.

New perturbative method improves stochastic gradient descent for binary weights.

problem Improving stochastic gradient descent for binary weights.
method Perturbative expansion around the mean of the sampling distribution, Taylor-corrected estimators, variance reduction techniques.
result Perturbative correction improves convergence of stochastic variational inference.

The paper analyzes the Heath-Jarrow-Morton-Musiela equation with Lévy perturbation.

problem Analyzing the Heath-Jarrow-Morton-Musiela equation with Lévy perturbation in weighted spaces.
method The paper studies the Heath-Jarrow-Morton-Musiela equation in weighted spaces of functions defined on [0,+)[0,+\infty), obtaining sufficient conditions for local and global existence.
result Conditions for global existence are close to necessary conditions for the linear diffusion term.

New methods protect malware classification networks from adversarial attacks.

problem Adversarial perturbations compromise malware classification networks.
method Training restricted networks with non-negative weight restrictions and relaxing constraints.
result Improved classifier accuracy while maintaining resistance to adversarial attacks.

Most random ReLU networks are vulnerable to small, Euclidean adversarial perturbations.

problem Vulnerability of ReLU networks to adversarial attacks.
method Analysis of random ReLU networks with decreasing dimensions, using gradient flow and descent.
result Most examples can be perturbed by small Euclidean distances via gradient methods.

LoRA fine-tuning explained with gradient dynamics for low-rank perturbations.

problem Understanding why gradient descent converges to useful low-rank perturbations in LoRA fine-tuning.
method Generalized student-teacher setting with i.i.d. samples and online gradient descent.
result Gradient descent converges to the teacher model in dkO(1)dk^{O(1)} iterations under certain conditions.

Let ΩΩ be an open half-space or slab in Rn+1\mathbb{R}^{n+1} endowed with a perturbation of the Gaussian measure of the form f(p):=exp(ω(p)cp2)f(p):=\exp(ω(p)-c|p|^2), where c>0c>0 and ωω is a smooth concave function depending only on the signed distance from the linear hyperplane parallel to Ω\partialΩ. In this work we follow a varia…

2014-03-18abs ↗pdf ↗

Gradient-based methods introduce noise that penalizes models sensitive to weight perturbations.

problem Noise in gradient-based optimization methods.
method Analysis of Gradient Descent (GD) and Stochastic Gradient Descent (SGD) updating all parameters simultaneously.
result Noise introduced by simultaneous parameter updates penalizes models sensitive to weight perturbations.

New metrics improve scRNA-seq perturbation modeling by reducing mode collapse.

problem Outperformed by simple mean prediction in scRNA-seq perturbation modeling.
method Introduce DEG-aware metrics (WMSE, Rw2(Δ)R^{2}_{w}(Δ)) and negative/positive baselines.
result WMSE loss function reduces mode collapse and improves model performance.

Batch normalization makes neural networks more vulnerable to small adversarial perturbations.

problem Adversarial vulnerability of neural networks trained with batch normalization.
method Investigated the impact of batch normalization on adversarial robustness and compared it to weight decay.
result Substituting weight decay for batch norm nullifies adversarial vulnerability.

New neural networks without weight transport are more robust to adversarial attacks.

problem Existing neural networks are easily fooled by adversarial attacks.
method Trained neural networks using feedback alignment instead of weight transport.
result Neural networks without weight transport are significantly harder to fool (98% adversarial accuracy vs 0.03%).

Adversarial training purifies hidden weights to remove small perturbations.

problem Understanding and removing adversarial perturbations in deep learning models.
method Introducing Feature Purification, a principle that adversarial training aims to remove small dense mixtures in hidden weights.
result Adversarial training can make neural networks robust against small perturbations, even with simple algorithms.

Paper proposes WD-DP ERM for distributed learning with improved privacy and performance.

problem Training models in distributed settings with privacy and performance guarantees.
method Weighted distributed differential privacy (WD-DP) for ERM, considering different weights of clients.
result Improved noise bound and excess empirical risk bound in distributed settings.

RSO uses random weight perturbations to train deep networks without gradients.

problem Training deep neural networks efficiently and without gradient information.
method RSO is a gradient-free Markov Chain Monte Carlo approach that updates weights based on mini-batch loss reduction.
result RSO achieves high accuracy (99.1% on MNIST) with significantly fewer updates than traditional methods.

Spectral norm regularization improves deep learning models' generalizability.

problem High sensitivity to input perturbation degrades deep learning model performance.
method Spectral norm regularization, penalizing high spectral norm of weight matrices.
result Models trained with spectral norm regularization show better generalizability.

EMAP finds minimal perturbations to change model predictions, combining feature weighting and counterfactuals.

problem Improving model explanations for black box classifiers.
method Neural network approach that returns minimal adversarial perturbations.
result EMAP provides more interpretable explanations and faster than sampling-based methods.

New adversarial training methods generate multiplicative perturbations for robust DNN training.

problem Training Deep Neural Networks with adversarial examples to improve robustness.
method Proposes xAT and xVAT, generating multiplicative perturbations for robust training.
result xAT and xVAT match or outperform state-of-the-art classification accuracies and are faster.

Study shows OWL-regularized regression, including OSCAR, is vulnerable to adversarial perturbations.

problem Vulnerability of sparse regression models with strongly correlated covariates to adversarial perturbations.
method Formulated adversarial attack as an optimization problem and analyzed OSCAR's robustness.
result Regression performance of grouping strongly correlated features can be severely degraded under adversarial settings.

Improved neural network convergence with causal Bayesian modeling in retail performance.

problem Improving neural network convergence in retail performance models.
method Causal Bayesian neural network implementation, removal of weakest SEM path, Flipout layers, Vadam optimizer.
result Neural network convergence improved with removal of the weakest SEM path.

Differentially private method for estimating individualized treatment rules.

problem Estimating individualized treatment rules while preserving privacy.
method Differentially private two-stage empirical risk minimization (DP-2ERM).
result Improved privacy-utility trade-off demonstrated through simulations and applications.

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

problem Real-world perturbations are not well characterized in adversarial defenses.
method A conditional generator defines perturbation sets over latent space, with properties for quality measured.
result Learned perturbation sets generate diverse, meaningful perturbations and improve model robustness.

The paper generalizes K-stability results to singular and weighted settings.

problem Generalizing K-stability to singular and weighted settings.
method Generalization of results in \cite{Li22a} to singular and weighted settings.
result The \(\mathbb{G}\)-uniform weighted K-stability for models implies \(\mathbb{G}\)-coercivity of the weighted Mabuchi functional.

TULiP estimates uncertainty for deep learning models safely.

problem Reliable uncertainty estimation for deep learning models in the open world.
method TULiP considers a hypothetical perturbation, bounds its effect, and computes uncertainty from sampled predictions.
result TULiP achieves state-of-the-art performance in OOD detection benchmarks.

Study analyzes perturbations in singular subspaces under random noise.

problem Understanding singular vector and subspace changes in signal-plus-noise models.
method Generalized Davis-Kahan-Wedin theorem for any unitarily invariant norm, considering \ell_\infty and 2,\ell_{2,\infty} bounds.
result Fine-grained insights into singular vector and subspace perturbations, including \ell_\infty and 2,\ell_{2,\infty} bounds.

Construct minimal Lagrangian surfaces in complex projective plane via loop group method.

problem Construct minimal Lagrangian immersions from arbitrary Riemann surfaces into complex projective plane.
method Loop group method, perturbed equivariant minimal Lagrangian surfaces, Delaunay cylinders approximation.
result Construct a class of minimal Lagrangian cylinders approximating Delaunay cylinders.

A new PLL method uses class activation values to improve robustness.

problem Weakly supervised learning with noisy data and adversarial perturbations.
method Subjective logic with class activation values for uncertainty representation and label weight re-distribution.
result More robust predictions under high noise levels, out-of-distribution examples, and adversarial perturbations.

The paper provides robustness bounds for manifold learning techniques.

problem Understanding the robustness of manifold learning methods.
method Derives perturbation bounds for Procrustes, Classical Scaling, and Trilateration.
result Performance bounds for Isomap, Landmark Isomap, and Maximum Variance Unfolding are derived.

The study analyzes prediction errors in systems with memory kernels, providing bounds and stability results.

problem Prediction errors in stochastic dynamical systems with memory kernels.
method Analysis of generalized Langevin equations (GLEs) with Volterra equations, integrating synchronized noise coupling and weighted norms.
result Prediction discrepancies decay at a rate determined by the memory kernel's decay, quantitatively bounded by kernel estimation errors.

Unified PAC-Bayesian framework for deep learning generalization.

problem Limitations of existing PAC-Bayesian norm-based bounds for deep neural networks.
method Unified framework using anisotropic Gaussian posteriors and sensitivity matrix.
result Comparable or tighter generalization bounds compared to state-of-the-art approaches.

Study optimal strategies for insurer's dividends, investments, and liabilities.

problem Maximize insurer's utility of dividend payments over an infinite horizon.
method Perturbation approach to obtain optimal strategy and value function in closed form.
result Obtained optimal strategy and value function for log and power utility.