Study on adversarial robustness in neural networks across initialization and training phases.
problem Understanding adversarial robustness in neural networks during different learning stages.
method Analyzes adversarial robustness in various scenarios of over-parameterized networks with quadratic targets and infinite samples.
result Robustness can worsen when test error improves, and vice versa, revealing new tradeoffs.
This paper examines weight initialization for 1-Lipschitz networks to improve robustness against adversarial attacks.
problem Improving the robustness of deep neural networks against adversarial attacks.
method Examined weight parametrization of AOL and SLL networks, calculated weight variance bounds, and demonstrated weight decay.
result Weight initialization causes deep 1-Lipschitz networks to decay to zero, and weight variance does not affect output variance distribution.
Paper presents robust clustering methods for general mixture models.
problem Clustering with sub-Gaussian error assumptions often invalid in practice.
method Hybrid clustering with robust centroid estimate and data-driven initialization.
result Provably near-optimal mislabeling guarantees for general error distributions.
FCNv2 robustness tested under noise and random initial conditions.
problem Assessing AI weather forecasting model robustness to input noise.
method Two experiments with varying noise levels and random initial conditions.
result FCNv2 preserves hurricane features under low to moderate noise, but underestimates intensity and persistence.
Genus one singularity appears in mean curvature flow for certain initial conditions.
problem Understanding singularities in mean curvature flow.
method Analyzing one-parameter families of initial conditions in R3. result A robust genus one singularity appears in mean curvature flow.
New method improves adversarial training efficiency and robustness.
problem High computational costs and lack of stability in adversarial training.
method Backward smoothing for randomized smoothing of random initialization.
result Our method achieves similar model robustness as state-of-the-art methods but with significantly less training time.
HOPE improves SSMs for long-memory tasks with robust initialization and training.
problem Improving state-space models for long-memory tasks with robust initialization and training.
method Developed a new parameterization scheme called HOPE using Hankel operators and Markov parameters.
result HOPE improves SSMs' performance on Long-Range Arena tasks and demonstrates non-decaying memory.
Residual networks (ResNet) and weight normalization play an important role in various deep learning applications. However, parameter initialization strategies have not been studied previously for weight normalized networks and, in practice, initialization methods designed for un-normalized networks are used as a proxy.…
Understanding deep neural networks is a major research objective with notable experimental and theoretical attention in recent years. The practical success of excessively large networks underscores the need for better theoretical analyses and justifications. In this paper we focus on layer-wise functional structure and…
New method improves robustness in partially observable domains by training against latent distribution shifts.
problem Challenges in robustness under latent distribution shift in partially observable reinforcement learning.
method Formalizes adversarial latent-initial-state POMDP, proves minimax principle, derives best-response inequalities.
result Reduces robustness gaps from 10.3 to 3.1 shots with targeted exposure to shifted latent distributions.
Paper shows robustness of gradient descent in matrix sensing despite perturbations.
problem Understanding robustness of gradient descent in matrix sensing.
method Developed perturbed gradient flow to capture noise and improve robustness.
result Gradient descent is robust to perturbations in matrix sensing.
Study spectral learning for odeco tensors, addressing initialization bottlenecks.
problem Recovering orthogonally decomposable tensors under noise.
method Investigates perturbation bounds, non-convex optimization, and initialization strategies.
result Initialization is the main bottleneck for efficient algorithms.
Improved portfolio optimization reduces sensitivity to neural network initialization.
problem High sensitivity to neural network initialization in portfolio optimization.
method Robust end-to-end framework for risk budgeting portfolios.
result Enhanced stability in portfolio optimization without compromising performance.
New method for robust fixed-point smoothing without state augmentation.
problem Estimating initial states in Gaussian smoothing algorithms.
method Cholesky-based formulation without state augmentation.
result Matches runtime and robustness of existing methods.
Global convergence for robust regression problems via IRLS with enhancements.
problem Global convergence for robust regression problems.
method Augmentations to IRLS to ensure global recovery and improved robustness.
result Global recovery guarantees for robust regression problems, outperforming state-of-the-art algorithms.
Simple AMP algorithm robust to adversarial corruption.
problem Robust approximate message passing in spiked matrix models.
method Spectral pre-processing combined with robust spectral initialization.
result AMP output is close to correct for corrupted data.
New method tackles convergence issues in approximating FBSDEs.
problem Convergence issues in approximating coupled FBSDEs.
method Approximates initial condition for a family of FBSDEs, then uses it to approximate the original FBSDE.
result Method converges even when standard deep BSDE method fails.
A framework is presented for unsupervised learning of representations based on infomax principle for large-scale neural populations. We use an asymptotic approximation to the Shannon's mutual information for a large neural population to demonstrate that a good initial approximation to the global information-theoretic o…
This paper improves flow models to better handle perturbations in real-world data.
problem Flow models amplify initial errors in perturbed data, leading to poor generalization.
method Utilizes Bernstein-type polynomials to construct Normalizing Flows (NF) for higher robustness.
result Proposed NF framework provides theoretical upper bounds and practical advantages.
This paper solves tensor robust principal component analysis via scaled gradient descent.
problem Extracting useful information from tensor data robust to corruptions and ill-conditioning.
method Directly recovers low-rank tensor factors via scaled gradient descent with adaptive thresholding.
result The proposed algorithm converges linearly to the true low-rank tensor at a constant rate independent of the condition number.
New RL method tackles sim-to-real gap using interactive data collection.
problem Sim-to-real gap in reinforcement learning.
method Distributionally robust reinforcement learning with interactive data collection.
result Proves sample-efficient learning is impossible without additional assumptions.
Randomly initialized ReLU networks of depth two can approximate smooth functions well.
problem Approximation power of two-layer networks of random ReLUs.
method Harmonic analysis and ridgelet representation theory for upper bounds, dimensionality arguments for lower bounds.
result Near-matching upper and lower bounds for L2-approximation and Sobolev norms. Regression mixture models are widely studied in statistics, machine learning and data analysis. Fitting regression mixtures is challenging and is usually performed by maximum likelihood by using the expectation-maximization (EM) algorithm. However, it is well-known that the initialization is crucial for EM. If the init…
New method improves deep learning model robustness and accuracy for long sequences.
problem Challenges in learning long-range sequence tasks using state-space models.
method Proposes a perturb-then-diagonalize (PTD) methodology to address ill-posed diagonalization problems in SSMs.
result Demonstrates improved robustness and accuracy of S5-PTD model on Long-Range Arena benchmark.
Theoretical guarantees for STE, a robust subspace recovery method.
problem Recovering a low-dimensional subspace from corrupted data.
method Subspace-constrained Tyler's estimator (STE) with initialization conditions.
result STE can effectively recover the subspace under certain conditions.
Study on proper learning under relaxed worst-case robust loss for VC classes.
problem Proper adversarially robust PAC learning under relaxed worst-case robust loss.
method Introduced a family of robust loss relaxations and showed their effectiveness for proper learnability.
result VC classes are properly PAC learnable with sample complexity close to standard PAC learning setup.
Modern neural networks are typically trained in an over-parameterized regime where the parameters of the model far exceed the size of the training data. Such neural networks in principle have the capacity to (over)fit any set of labels including pure noise. Despite this, somewhat paradoxically, neural network models tr…
New hyperparameter ensembles boost neural network performance and uncertainty.
problem Improving neural network robustness and uncertainty quantification.
method Designing ensembles over both weights and hyperparameters, stratified across random initializations.
result Hyper-deep and hyper-batch ensembles outperform deep and batch ensembles on various architectures.
Fine-tuning through knowledge transfer from a pre-trained model on a large-scale dataset is a widely spread approach to effectively build models on small-scale datasets. In this work, we show that a recent adversarial attack designed for transfer learning via re-training the last linear layer can successfully deceive m…
We consider the martingale optimal transport duality for càdlàg processes with given initial and terminal laws. Strong duality and existence of dual optimizers (robust semi-static superhedging strategies) are proved for a class of payoffs that includes American, Asian, Bermudan, and European options with intermediate m…
Robust deep neural networks estimate multi-dimensional functional data robustly.
problem Estimating location function from multi-dimensional functional data robustly.
method Deep neural networks with ReLU activation, robust to outliers and model misspecification.
result Uniform convergence rates for robust deep neural network estimators.
Model-based clustering approaches concern the paradigm of exploratory data analysis relying on the finite mixture model to automatically find a latent structure governing observed data. They are one of the most popular and successful approaches in cluster analysis. The mixture density estimation is generally performed …
Method for initializing Gaussian mixtures for variational inference with multi-modal distributions.
problem Challenges in variational inference with Gaussian mixtures due to multimodality and nonconvex loss functions.
method Optimization to find local maxima, local Gaussian approximations, and constrained least squares regression.
result Robust initialization improves variational inference performance and scalability.
Improves natural accuracy of deep learning models by combining robust predictions and features.
problem Maintaining natural accuracy while resisting adversarial attacks.
method Ensemble methods combining robust and standard models.
result Optimized natural accuracy through ensemble of robust models.
In this work, we investigate black-box optimization from the perspective of frequentist kernel methods. We propose a novel batch optimization algorithm, which jointly maximizes the acquisition function and select points from a whole batch in a holistic way. Theoretically, we derive regret bounds for both the noise-free…
Develops robust learning framework under distributional perturbations.
problem Learning robust to data distributional changes.
method Distributionally Robust Optimization (DRO) under Wasserstein metric.
result Establishes performance guarantees and tractable formulations.
Adversarially robust models transfer better than standard models in image classification.
problem Improving transfer learning performance in image classification.
method Focused on adversarially robust ImageNet classifiers, compared to standard models.
result Adversarially robust models yield improved accuracy on downstream classification tasks.
The performance of gradient-based optimization strategies depends heavily on the initial weights of the parametric model. Recent works show that there exist weight initializations from which optimization procedures can find the task-specific parameters faster than from uniformly random initializations and that such a w…
A new method averages neural network parameters to rank features robustly.
problem Neural networks' sensitivity to random initialization affects feature ranking robustness.
method Parameter averaging of multiple shallow networks trained with different random seeds.
result The averaged model discovers ground-truth feature importance consistently.
New method trains deep networks robustly without adaptive methods.
problem Training deep networks with robustness and efficiency.
method Scale invariant architecture + SGD + weight decay + gradient clipping.
result SGD can achieve similar performance to adaptive methods like Adam.
Study shows how pretraining robustness transfers to downstream tasks.
problem Understanding how robustness is transferred from pretraining to downstream tasks.
method Theoretical analysis and practical validation of robustness constraints.
result Robustness of a linear predictor on downstream tasks can be constrained by the robustness of its underlying representation.
Paper proposes faster certified robust training methods with short warmup.
problem Certified robust training methods require long warmup schedules, making training costly.
method Proposes three improvements: new weight initialization, BN, and regularization.
result Achieves 65.03% verified error on CIFAR-10 and 82.36% on TinyImageNet with short warmup.
Gradient boosting algorithms construct a regression predictor using a linear combination of ``base learners''. Boosting also offers an approach to obtaining robust non-parametric regression estimators that are scalable to applications with many explanatory variables. The robust boosting algorithm is based on a two-stag…
This work studies adversarial transferability and proposes ensemble methods to improve robustness.
problem Adversarial transferability in neural networks and its implications for robustness.
method Investigates the effect of various factors on adversarial transferability and proposes ensemble attack methods.
result Transferability is significantly hampered by input quantization and architectural mismatch, but not by initialization.
Abstract: Nonlinear random walk with distributionally robust transition probabilities.
problem Modeling nonlinear random walks with robust transition probabilities.
method Scaling limit and nonlinear semigroup approach.
result Explicit computation of the generator and corresponding PDE.
Improved training for VQ-VAE models with robust codebook learning.
problem Challenges in training discrete latent variable models, especially VQ-VAEs.
method Increased learning rate and periodic re-initialization of codebook for robust training.
result More robust training and increased usage of latent codewords, even for large codebooks.
Motivated by the gap between theoretical optimal approximation rates of deep neural networks (DNNs) and the accuracy realized in practice, we seek to improve the training of DNNs. The adoption of an adaptive basis viewpoint of DNNs leads to novel initializations and a hybrid least squares/gradient descent optimizer. We…
Robust Lasso-Zero handles missing covariates and sparse corruptions.
problem Sparse corruptions and missing covariates in sparse linear models.
method Extension of Lasso-Zero to handle sparse corruptions, with theoretical guarantees on sign recovery.
result Robust Lasso-Zero can handle missing values without specifying a parametric model.