BEGAN-CS prevents mode collapse in GANs by adding a latent-space constraint.
problem Mode collapse in BEGAN during training.
method Introducing a latent-space constraint in the loss function of BEGAN.
result BEGAN-CS improves training stability and suppresses mode collapse.
We propose a new equilibrium enforcing method paired with a loss derived from the Wasserstein distance for training auto-encoder based Generative Adversarial Networks. This method balances the generator and discriminator during training. Additionally, it provides a new approximate convergence measure, fast and stable t…
Faster reconstruction of compressed signals using conditional GAN and NPGD.
problem Recovering compressed signals from measurements.
method Network-based projected gradient descent (NPGD) combined with measurement-conditional generative adversarial networks (GANs/BEGANs).
result Significant speed-up in reconstruction (up to 140-175 times faster).
Paper analyzes adversarial attacks and defenses using game theory.
problem Unclear conditions for optimal attacks and defenses in adversarial learning.
method Game-theoretic framework with locally linear decision boundary model.
result Fast Gradient Method attack and Randomized Smoothing defense form a Nash Equilibrium.
Improves GANs training through game theory.
problem Hard training of GANs due to antagonistic networks.
method Rewrote GAN training as a variational inequality and introduced a stochastic relaxed forward-backward algorithm.
result Algorithm converges to an exact solution or a neighborhood of it under monotonicity.
Deep neural networks infer downlink CSI from uplink CSI without feedback.
problem Efficient allocation of wireless resources in FDD systems requires accurate DL-CSI.
method Used convolutional neural networks and GANs to infer DL-CSI from UL-CSI.
result Deep learning can accurately predict DL-CSI from UL-CSI for various environments.
We show that training of generative adversarial network (GAN) may not have good generalization properties; e.g., training may appear successful but the trained distribution may be far from target distribution in standard metrics. However, generalization does occur for a weaker metric called neural net distance. It is a…
New approaches improve adversarial robustness of DEQs.
problem Adversarial vulnerability of DEQs.
method Developed approaches to estimate intermediate gradients and integrate them into attacking pipelines.
result Demonstrated adversarial robustness of DEQs competitive with deep networks.
We propose a new algorithm for training generative adversarial networks that jointly learns latent codes for both identities (e.g. individual humans) and observations (e.g. specific photographs). By fixing the identity portion of the latent codes, we can generate diverse images of the same subject, and by fixing the ob…
AROS uses Lyapunov-stabilized embeddings to improve out-of-distribution detection robustness against adversarial attacks.
problem Robust out-of-distribution (OOD) detection against adversarial attacks.
method Neural Ordinary Differential Equations (NODEs) with Lyapunov stability theory for generating robust embeddings.
result Improves robust detection performance significantly, e.g., from 37.8% to 80.1% on CIFAR-10 vs. CIFAR-100.
New method UADs improves transferability of adversarial perturbations.
problem Transferability of adversarial perturbations across different DNN architectures.
method Proposes Universal Adversarial Directions (UADs) to improve transferability.
result UADs can achieve a Nash equilibrium, indicating potential transferability.
Save for some special cases, current training methods for Generative Adversarial Networks (GANs) are at best guaranteed to converge to a `local Nash equilibrium` (LNE). Such LNEs, however, can be arbitrarily far from an actual Nash equilibrium (NE), which implies that there are no guarantees on the quality of the found…
New theorem guarantees approximate equilibrium in non-convex games.
problem No guarantee of equilibrium in non-convex games.
method Introduced a minimax theorem for non-convex games involving neural networks.
result Provided an approximate minimax theorem for non-convex games.
Optimal adversarial noise algorithms for multiple classifiers using game theory.
problem Designing robust attacks against multiple classifiers.
method Formulating the problem as a two-player, zero-sum game and using Multiplicative Weights Update framework with best response oracles.
result Demonstrated the effectiveness of randomization in adversarial attacks and optimal mixed strategies.
Enhances deep networks' robustness against adversarial attacks.
problem Vulnerability of deep neural networks to adversarial examples.
method Boundary Conditional GAN, generating boundary samples near decision boundary.
result Significant improvement in robustness against various adversarial attacks.
The paper examines Nash equilibrium in GANs for stationary Gaussian processes.
problem Existence and uniqueness of Nash equilibrium in GANs for stationary Gaussian processes.
method Analyzes the existence of Nash equilibrium in GANs for stationary Gaussian processes, considering different discriminator families.
result The existence of Nash equilibrium depends on the discriminator family and symmetry properties of the generator family.
In this work, we consider the task of classifying binary positive-unlabeled (PU) data. The existing discriminative learning based PU models attempt to seek an optimal reweighting strategy for U data, so that a decent decision boundary can be found. However, given limited P data, the conventional PU models tend to suffe…
LOGAN optimizes GAN training by improving adversarial dynamics.
problem Training GANs is challenging due to delicate adversarial dynamics and potential divergence.
method Integrates natural gradient-based latent optimisation into CS-GAN.
result Significant improvement in GAN training performance, achieving state-of-the-art results.
New parameterization of neural networks with Lipschitz bounds for robustness.
problem Developing robust neural networks with Lipschitz bounds.
method Introducing a new parameterization that admits a Lipschitz bound during training without requiring projections or barrier functions.
result The new parameterization improves robustness to adversarial attacks in image classification.
Improves adversarial robustness of DEQ models by regulating neural dynamics.
problem Limited adversarial robustness of DEQ models.
method Interprets DEQs as neural dynamics, uses entropy reduction and random intermediate states.
result Significantly increases adversarial robustness of DEQ models.
Paper presents IPVI for unbiased DGP inference with efficiency.
problem Intractable exact inference in deep Gaussian processes.
method Implicit posterior variational inference framework.
result IPVI achieves unbiased posterior belief with efficiency.
New algorithm finds mixed Nash equilibria in GANs.
problem No provably convergent algorithm exists for general GANs.
method Developed a novel algorithmic framework via an infinite-dimensional two-player game.
result Proved rigorous convergence rates to the mixed Nash Equilibrium.
Generative model solves financial market equilibria with stable reinforcement learning.
problem Financial market equilibria under realistic frictions and multiple agents.
method Generative adversarial reinforcement learning with decoupling feedback.
result Algorithm learns and predicts asset returns and volatilities.
The study reveals how adversarial perturbations can include class features for generalization.
problem Understanding why adversarial examples deceive neural networks and transfer between networks.
method A one-hidden-layer network trained on mutually orthogonal samples.
result Adversarial perturbations, even of a few pixels, contain sufficient class features for generalization.
New technique helps GANs reach equilibrium by moving 'across' the curl.
problem Achieving equilibrium in GANs using gradient descent.
method Using Variational Inequalities to analyze GAN training algorithms.
result Convergence to equilibrium achieved through a specific orthogonal direction.
The paper identifies potential adversarial samples near decision boundaries of neural networks.
problem Vulnerability of deep neural networks to small perturbations of inputs.
method Developed a method to explore near decision boundaries of trained classifiers to identify potential adversarial samples.
result Potential adversarial samples represent only 61% of the test data but cover more than 82% of adversarial samples produced by iFGSM and 92% of those by DeepFool on CIFAR10.
Study introduces statistical mechanics for min-max problems.
problem Understanding the properties of min-max problems in high dimensions.
method Statistical mechanical formalism for analyzing min-max problems.
result Derives the relationship between training data and generalization error.
Gradient methods converge exponentially in concave network games.
problem Finding Nash equilibria in concave network zero-sum games.
method Gradient Ascent and Optimistic Gradient Ascent analyses.
result Exponential convergence rates in various game settings.
New model shows online and statistical learning are computationally equivalent with optimization oracle.
problem Online learning in non-convex games with adversarial settings.
method Strengthening the oracle model to make online and statistical learning computationally equivalent.
result Efficient computation of non-convex game equilibria, including GANs, with optimization oracle.
Researchers compute exact decision boundaries of neural networks.
problem Finding exact decision boundaries of trained neural networks is intractable.
method Developed mathematical tools to investigate decision boundaries of trained deep models.
result Some simplifying assumptions about decision boundaries are unreliable.
GANs may not have Nash equilibria, but proximal training can find solutions.
problem Existence of Nash equilibria in GANs optimization.
method Proximal training approach to find solutions.
result Proximal training finds solutions to GAN problems.
New GAN formulation addresses mode collapse issue.
problem Mode collapse in GANs.
method Randomized decision rules, empirical Bayes, stochastic gradient MCMC.
result Proposed method converges to Nash equilibrium.
This paper investigates how dataset properties affect GAN training outcomes.
problem GANs often fail to reach equilibrium due to instability or mode collapse.
method Experiments to identify patterns in dataset properties and their effects on GAN training.
result Patterns in dataset properties influence GAN training dynamics and outcomes.
Despite the considerable success enjoyed by machine learning techniques in practice, numerous studies demonstrated that many approaches are vulnerable to attacks. An important class of such attacks involves adversaries changing features at test time to cause incorrect predictions. Previous investigations of this proble…
Improved neural network robustness with instance-specific perturbation margins.
problem Adversarial training fails to generalize well to unperturbed test set.
method Instance adaptive adversarial training with sample-specific perturbation margins.
result Test accuracy improves with a marginal drop in robustness.
DeepDIG generates samples near decision boundaries of deep neural networks for better understanding.
problem Limited knowledge of how deep neural networks make decisions.
method Adversarial example generation to create samples near decision boundaries.
result Characterized decision boundaries of various deep neural network models.
New framework crafts adversarial examples for neural networks.
problem Understanding and defending against adversarial examples.
method Adversarial Example Games (AEG) framework for crafting transferable adversarial examples.
result AEG provides a theoretical foundation for crafting adversarial examples and guarantees their transferability.
New method generates equilibrium glass configurations efficiently.
problem Sampling equilibrium configurations of amorphous materials is slow and difficult.
method Riemannian stochastic interpolation framework combining Riemannian stochastic interpolant and equivariant flow matching.
result Enforcing geometric and symmetry constraints significantly improves generative performance.
Deep networks can overfit benignly but still be vulnerable to adversarial attacks.
problem Adversarial vulnerability of deep neural networks trained with benign overfitting.
method Investigated causes of adversarial vulnerability, identified label noise as a key factor, and explored the impact of training procedures and representation learning.
result Adversarial robustness requires more complex decision boundaries than simple ones, suggesting the need for better representation learning.
This paper examines various definitions of adversarial risk and their implications.
problem Quantifying the performance of classifiers under adversarial perturbations.
method Optimal transport, robust statistics, functional analysis, and game theory.
result Generalization of Strassen's theorem and new connections to Choquet capacities and game theory.
A new autoencoder uses stochastic functions to encourage diversity in generated samples.
problem Generating diverse samples from autoencoders.
method Replacing the adversary in AAE with a space of stochastic functions.
result More diverse set of generated samples.
This paper examines how adversarial perturbations affect model performance and equilibrium learning.
problem Adversarial perturbations and covariate shifts impact model performance and equilibrium learning.
method Characterizes the extrapolation region in regression and classification, analyzes dynamics of adversarial learning games.
result Establishes two directional convergence results: a blessing in regression and a curse in classification.
Two new methods improve causal inference without noise assumptions.
problem Improving causal inference without distributional assumptions.
method Adversarial Orthogonal Regression (AdOR) and AdOSE methods.
result Effective in synthetic and real-world experiments.
Study p-Willmore disks with boundary energies, finding equilibrium configurations.
problem Finding equilibrium configurations for p-Willmore disks with boundary energies.
method Model boundary as Kirchhoff elastic rod, interior term dependent on mean and Gaussian curvatures. Study among topological disks and p-Willmore examples.
result Equilibrium configurations for p-Willmore disks with boundary energies.
Paper presents an algorithm to analyze DGNNs by identifying generative boundaries and sampling.
problem Challenges in identifying the generation mechanism of DGNNs.
method Explorative sampling algorithm to analyze DGNNs by defining generative boundaries and optimizing essential set.
result Empirically reveals the characteristics of DGNNs' internal layers and finds more homogeneous samples.
This work uses tropical geometry to understand neural network decision boundaries.
problem Characterizing neural network decision boundaries with piecewise linear activations.
method Tropical geometry applied to a simple neural network model.
result Decision boundaries are a subset of a tropical hypersurface related to a polytope formed by zonotopes.
A novel method to train networks from each other's adversarial examples to resist black-box attacks.
problem Machine learning models can be fooled by adversarial examples, especially in black-box attacks.
method Simultaneous adversarial training combining two networks to learn from each other's adversarial examples.
result The method improves the networks' resilience to black-box attacks.
Paper tackles online allocation problems using adversarial training.
problem Online bipartite matching, especially in AdWords.
method Constructs a framework combining game theory, adversarial training, and GANs.
result Designs robust algorithms that perform well under practical and adversarial conditions.