SAM selects flatter minima late in training, improving generalization.
problem Improving neural network generalization under various settings.
method Sharpness-Aware Minimization (SAM) applied late in training.
result SAM efficiently selects flatter minima late in training, improving generalization.
VL finds flatter solutions at edge of stability, matching theory with practice.
problem Understanding implicit regularization in deep learning.
method Edge of Stability framework, controlling variational posterior shape and sample number.
result VL finds even flatter solutions than gradient descent.
GD converges faster to flatter minima than gradient flow in shallow networks.
problem Understanding the dynamics of gradient descent in shallow linear networks.
method Analyzing the convergence rate and solution of gradient descent in depth-2 linear neural networks.
result GD converges linearly to flatter minima than gradient flow, even with large step sizes.
In this work we study generalization of neural networks in gradient-based meta-learning by analyzing various properties of the objective landscapes. We experimentally demonstrate that as meta-training progresses, the meta-test solutions, obtained after adapting the meta-train solution of the model, to new tasks via few…
Gradient descent with large momentum finds flatter minima.
problem Understanding the effects of momentum in gradient descent.
method Empirical and theoretical analysis of gradient descent with large momentum.
result Large momentum leads to flatter minima than gradient descent.
Study shows deep linear networks can converge to flatter minima at large learning rates.
problem Understanding the implicit bias of deep linear networks at large learning rates.
method Characterization of deep linear networks for binary classification using logistic loss in the large learning rate regime.
result Gradient descent iterates converge to a flatter minimum in the catapult phase for certain data separation conditions.
mSAM improves generalization by making models flatter.
problem Over-parameterized models can have varying generalization performance.
method mSAM modifies loss function to favor flatter minima through adversarial perturbations.
result mSAM achieves flatter minima than SAM and SGD, improving generalization.
Expressive quantum circuits are harder to train due to flatter cost landscapes.
problem Designing quantum circuits that are both expressive and trainable.
method Deriving a relationship between expressibility and gradient magnitude, extending barren plateau phenomenon.
result Highly expressive ansätze exhibit flatter cost landscapes, making them harder to train.
Quantum neural networks generalize better due to flatter parameter space.
problem Generalization in quantum neural networks.
method Mapped feature data to a quantum state, applied unitary evolution, and measured for classification.
result Quantum neural networks have better generalization than classical networks.
The pursuit of explaining and improving generalization in deep learning has elicited efforts both in regularization techniques as well as visualization techniques of the loss surface geometry. The latter is related to the intuition prevalent in the community that flatter local optima leads to lower generalization error…
In several experimental reports on nonconvex optimization problems in machine learning, stochastic gradient descent (SGD) was observed to prefer minimizers with flat basins in comparison to more deterministic methods, yet there is very little rigorous understanding of this phenomenon. In fact, the lack of such work has…
SALR improves deep learning generalization by dynamically adjusting learning rates.
problem Improving generalization in deep learning models.
method Sharpness-aware learning rate scheduling based on local loss function sharpness.
result SALR drives solutions to flatter regions, improving generalization and convergence.
Noise in RNNs promotes flatter minima and more stable dynamics.
problem Understanding and optimizing the training of RNNs with noise.
method Formalizing RNNs as stochastic differential equations and analyzing the effect of noise in the hidden states.
result Noise injection in RNNs leads to flatter minima, more stable dynamics, and improved robustness.
Deep neural networks are typically trained by optimizing a loss function with an SGD variant, in conjunction with a decaying learning rate, until convergence. We show that simple averaging of multiple points along the trajectory of SGD, with a cyclical or constant learning rate, leads to better generalization than conv…
The paper analyzes how noise geometry influences the performance of SGD in machine learning.
problem Understanding how noise geometry affects the performance of stochastic gradient descent.
method Developed two metrics to quantify noise alignment strength and analyzed their effects on loss and subspace projection dynamics.
result Noise geometry can be used to guarantee alignment under certain conditions, aiding SGD's ability to escape from sharp minima.
This work analyzes statistical properties of SAM, showing it outperforms GD.
problem Improving deep neural network generalization through flatter solutions.
method Directly studies statistical performance of Sharpness-Aware Minimization (SAM).
result SAM has smaller prediction error than Gradient Descent (GD) under certain conditions.
New model shows SGD can prefer sharp or flat solutions based on label noise.
problem Understanding SGD's preference for flat or sharp solutions during training.
method Solved an analytically solvable model to explore SGD behavior.
result Data distribution determines sharpness at convergence; isotropic label noise leads to flat minimum preference.
It was empirically confirmed by Keskar et al.\cite{SharpMinima} that flatter minima generalize better. However, for the popular ReLU network, sharp minimum can also generalize well \cite{SharpMinimacan}. The conclusion demonstrates that the existing definitions of flatness fail to account for the complex geometry of Re…
New research shows flat minima in robust loss landscapes correlate with good adversarial robustness.
problem Adversarial training leads to robust overfitting, poor robust generalization.
method Average- and worst-case metrics to measure flatness in robust loss landscapes.
result Flatness in robust loss landscapes correlates with good adversarial robustness.
Function-space MAP estimation leads to better generalization and robustness.
problem The mismatch between parameter posterior and function posterior in model training.
method Directly estimating the most likely function implied by the model and data.
result Function-space MAP estimation can lead to flatter minima, better generalization, and improved robustness.
Momentum affects optimization differently at small vs large batch sizes near instability.
problem Understanding how momentum impacts optimization near the edge of stability.
method Demonstrated through batch-size dependent behavior of SGD with momentum.
result Momentum operates in two distinct regimes: amplifying stochastic fluctuations at small batch sizes and stabilizing at large batch sizes.
Twin-Boot integrates uncertainty estimation into optimization using parallel training of identical models.
problem Uncertainty in overparameterized models, especially in low-data regimes.
method Twin-Bootstrap Gradient Descent (Twin-Boot) trains two identical models on independent bootstrap samples and uses their divergence to guide learning.
result Improves calibration and generalization, yields interpretable uncertainty maps.
New findings cast doubt on the role of λmax in generalizing neural networks.
problem The role of λmax in neural network generalization remains unclear. method Experiments with various training interventions and batch sizes.
result Generalization benefits can vanish at larger batch sizes, challenging the role of λmax. Improved electrical load forecasting model using Fourier-enhanced RNN.
problem Electrical load time series downscaling with high accuracy and low error.
method Combines recurrent neural network with Fourier seasonal embeddings and self-attention.
result Significantly reduces RMSE across different time horizons compared to existing methods.
ε-shotgun optimizes expensive function evaluations in parallel.
problem Optimizing expensive black-box functions efficiently.
method ε-greedy batch Bayesian optimisation.
result ε-shotgun outperforms state-of-the-art batch methods.
SGD batch size affects autoencoder global minima sparsity and sharpness.
problem Investigating how batch size impacts autoencoder learning.
method Non-convex autoencoder training with SGD, varying batch sizes.
result SGD batch size influences global minimum sparsity and sharpness.
Truncated SGD with heavy-tailed noise eliminates sharp local minima.
problem Avoiding sharp local minima in deep learning models.
method Truncated SGD with heavy-tailed gradient noise.
result Truncated SGD can eliminate sharp local minima entirely from its training trajectory.
Hyperelastic bodies in Riemannian manifolds can levitate due to curvature-induced forces.
problem Hyperelastic bodies in Riemannian manifolds can levitate due to curvature-induced forces.
method Numerical simulations of static solutions to a particular class of problems in hyperelastic mechanics.
result Hyperelastic bodies in Riemannian manifolds can levitate due to curvature-induced forces.
Label noise SGD converges to a simple model with a single linear feature.
problem Understanding the simplicity bias in neural network training.
method Analyzing the convergence of label noise SGD on two-layer neural networks.
result Label noise SGD converges to a model with a single linear feature.
Stochastic gradient descent (SGD) is almost ubiquitously used for training non-convex optimization tasks. Recently, a hypothesis proposed by Keskar et al. [2017] that large batch methods tend to converge to sharp minimizers has received increasing attention. We theoretically justify this hypothesis by providing new pro…
Double descent found in DRL, improving generalization with model capacity.
problem Generalization in over-parameterized DRL models.
method Actor-Critic framework, Policy Entropy metric.
result Policy Entropy significantly reduces as model capacity increases, indicating improved generalization.
Sharpness minimization algorithms don't solely improve generalization.
problem Why do overparameterized neural networks generalize?
method Theoretical and empirical investigation of two-layer ReLU networks.
result Sharpness minimization algorithms do not always lead to better generalization.
Adaptive stepsizing improves sampling in Bayesian neural networks.
problem Scalable sampling of posterior distributions in Bayesian neural networks.
method SA-SGLD, employing time rescaling to adapt stepsize dynamically.
result SA-SGLD achieves more accurate posterior sampling than SGLD.
New method accelerates neural network training by focusing on flat directions.
problem Improving neural network training speed and stability.
method Bulk-SGD, interpolated gradient methods.
result Updates along the Dominant subspace can accelerate convergence but compromise stability.
New findings show mini-batch SGD operates in a 'Edge of Stochastic Stability' regime.
problem Understanding the stability and convergence of mini-batch SGD.
method Analyzing the mini-batch Hessian and its directional curvature.
result Mini-batch SGD operates in a different stability regime (Edge of Stochastic Stability) compared to full-batch GD.
Label noise in SGD helps converge to flatter minima.
problem Improving generalization in overparametrized models.
method Analyzes SGD with label noise, showing convergence to regularized minima.
result SGD with label noise converges to flatter minima, improving generalization.
Unlike human speakers, typical text-to-speech (TTS) systems are unable to produce multiple distinct renditions of a given sentence. This has previously been addressed by adding explicit external control. In contrast, generative models are able to capture a distribution over multiple renditions and thus produce varied r…
Transformers are less sensitive to input perturbations compared to other models.
problem Understanding the inductive biases of transformers and distinguishing them from other architectures.
method Identified token-wise sensitivity as a metric to explain transformers' inductive biases across different data modalities.
result Transformers have lower sensitivity than MLPs, CNNs, ConvMixers, and LSTMs, across vision and language tasks.
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.
With the industry trend of shifting from a traditional hierarchical approach to flatter management structure, crowdsourced performance assessment gained mainstream popularity. One fundamental challenge of crowdsourced performance assessment is the risks that personal interest can introduce distortions of facts, especia…
New model explains deep learning performance at large learning rates.
problem Understanding deep learning performance at different learning rates.
method Developed neural networks with solvable training dynamics.
result Large learning rates lead to convergence to flatter minima.
The local geometry of high dimensional neural network loss landscapes can both challenge our cherished theoretical intuitions as well as dramatically impact the practical success of neural network training. Indeed recent works have observed 4 striking local properties of neural loss landscapes on classification tasks: …
Large learning rates cause parameter instability, leading to better generalization.
problem Understanding why deep neural networks perform well despite operating outside the traditional stability regime.
method Analyzing the effect of large learning rates on the orientation of Hessian eigenvectors and parameter exploration.
result Large learning rates induce parameter instability, leading to better generalization through exploration of flatter regions of the loss landscape.
Deep neural networks trained on a wide range of datasets demonstrate impressive transferability. Deep features appear general in that they are applicable to many datasets and tasks. Such property is in prevalent use in real-world applications. A neural network pretrained on large datasets, such as ImageNet, can signifi…
Paper confirms winning tickets have sharp minima, useful for generalization.
problem Explaining why over-parameterized models generalize well despite having sharp minima.
method PAC-Bayesian theory applied to analyze winning tickets and their generalization behavior.
result PAC-Bayesian theory confirms winning tickets have relatively sharp minima, a disadvantage for generalization.
X-DER improves continual learning by revising replay memory and learning unseen classes.
problem Deep networks forget when learning new tasks sequentially.
method Combines rehearsal and Knowledge Distillation, revising replay memory and learning unseen classes.
result X-DER outperforms state-of-the-art on various benchmarks.
Paper tackles privacy-preserving data density issues using deconvolution.
problem Privacy-preserving noise affects data density, leading to under/over-estimation.
method Develops deconvoluting kernel density estimators and regression models.
result Demonstrates improved accuracy in estimating heavy-hitters with locally differential data.
Adaptor 'E' extends gradient-based optimizers to explore loss landscapes, improving generalization.
problem Finding lower and better-generalizing minima in deep learning.
method Proposes an adaptor 'E' to extend gradient-based optimizers, encouraging exploration along landscape valleys.
result Adapted optimizers increase test accuracy by an average of 2.5% in large-batch training tasks.