Study loop corrections in random feature models affecting training and test errors.
problem Analyzing loop corrections in random feature models to understand training and test errors.
method Statistical physics and effective field theory approach to study loop corrections.
result Derived loop corrections to training error, test error, and generalization gap.
DD algorithm tracks test error from train error without validation data.
problem Systematic generalization gap between train and test errors in modern model training.
method Decoupled descent (DD) algorithm that cancels data reuse biases via approximate message passing.
result DD algorithm rigorously demonstrates zero-cost validation and 100% data utilization.
A form of generalisation error known as Off Training Set (OTS) error was recently introduced in [Wolpert, 1996b], along with a theorem showing that small training set error does not guarantee small OTS error, unless assumptions are made about the target function. Here it is shown that the applicability of this theorem …
New techniques improve distributed training with compressed gradients.
problem Gradient mismatch problem in local error feedback.
method Step-ahead error feedback and error averaging techniques.
result Our methods handle gradient mismatch and train faster than full-precision training.
Optimal machine learning requires interpolating training data in high-dimensional linear regression.
problem Achieving optimal predictive risk in overparameterized linear regression models.
method Analyzing proportional asymptotics of random design and label noise variance.
result Optimal performance in linear regression requires fitting training data to higher accuracy than inherent noise.
Study shows label errors impact model disparity metrics, proposing mitigation methods.
problem Impact of label errors on model disparity metrics.
method Empirical study, characterizing label error effects; proposing estimation and relabeling methods.
result Label errors significantly affect model disparity metrics, particularly for minority groups.
Curiosity-Critic improves world model training by focusing on cumulative prediction error.
problem Training world models with intrinsic rewards that consider cumulative prediction error.
method Curiosity-Critic uses a surrogate reward based on the difference between current and asymptotic prediction errors, estimated online by a co-trained critic.
result Curiosity-Critic outperforms other methods in training speed and final world model accuracy.
Adversarial training achieves optimal test error for shallow networks.
problem Achieving optimal adversarial test error for general data distributions.
method Applying new Rademacher complexity bounds and properties of optimal adversarial predictors.
result Adversarial training can achieve optimal adversarial test error for general data distributions.
Adversarial training augments the training set with perturbations to improve the robust error (over worst-case perturbations), but it often leads to an increase in the standard error (on unperturbed test inputs). Previous explanations for this tradeoff rely on the assumption that no predictor in the hypothesis class ha…
Study shows accuracy of neural networks depends more on error location than percentage of error.
problem Effect of noise on accuracy in incremental learning neural networks.
method Empirical study using Perceptron, Feed Forward Neural Network, and Radial Basis Function Neural Network.
result Accuracy of neural networks is more dependent on error location than the percentage of error.
We trained three Binarized Convolutional Neural Network architectures (LeNet-4, Network-In-Network, AlexNet) on a variety of datasets (MNIST, CIFAR-10, CIFAR-100, extended SVHN, ImageNet) using error-prone activations and tested them without errors to study the resilience of the training process. With the exception of …
Estimates the generalization error of deep neural networks without relying on capacity measures.
problem Understanding how generalization error scales with training data for deep neural networks.
method Derives estimates of generalization error for deep networks based on two assumptions: zero training error and error probability proportional to distance to nearest training point.
result Estimates the generalization error of DNNs as O(1/(δN^{1/d})), matching experimental behavior.
ECN framework improves training on noisy structured labels.
problem Structured errors in fine-grained annotations lead to biased models.
method Error-Correcting Networks (ECN) framework.
result ECN improves fine-grained annotation prediction.
Analyzes the generalization and training errors of the random feature model over time.
problem Understanding the temporal behavior of generalization and training errors in deep learning.
method Uses Cauchy complex integral representations and random matrix methods based on linear pencils.
result Analytical solution of the full time-evolution path of generalization and training errors.
Modified training direction reduces generalization error in neural networks.
problem Reducing generalization error in neural networks.
method Theoretical analysis of modified natural gradient descent in function space.
result Modifying training direction in function space reduces total generalization error.
Bias correction needed after deep learning regression training.
problem Systematic error accumulation in deep learning regression models.
method Adjust bias of the machine learning model post-training.
result Bias correction efficiently solves error accumulation.
Random forest training yields confidence intervals for generalization error.
problem Computing accurate confidence intervals for random forest generalization error.
method Directly computes confidence intervals from training data without data splitting.
result Confidence intervals have good coverage and appropriate width.
Paper analyzes Gibbs and Langevin Monte Carlo for interpolation regime, showing generalization from low errors.
problem Analyzing Gibbs and Langevin Monte Carlo in overparameterized interpolation regime.
method Data-dependent bounds and stability under approximation with Langevin Monte Carlo.
result Generalization is signaled by small training errors in noisy regime, with bounds stable under approximation.
Neural operators correct PDE residuals to improve BIP solutions.
problem Reducing error in infinite-dimensional Bayesian inverse problems with neural operators.
method Error correction using PDE residuals to improve neural operator approximation.
result Trained neural operators with error correction achieve a quadratic reduction in approximation error.
This paper provides a mathematical foundation for deep neural networks solving PDEs.
problem Mathematical foundation for deep neural networks solving high-dimensional PDEs.
method Decomposed generalization error into approximation and training errors; derived gradient flow in the wide network limit.
result Generalization error tends to zero as the number of neurons and training time tend to infinity.
The study provides a comprehensive error analysis for diffusion models.
problem Understanding the accuracy and design of diffusion-based generative models.
method Conducts a non-asymptotic convergence analysis and sampling error analysis for diffusion models.
result Theoretical insights into designing effective training and sampling processes for diffusion models.
SGD-trained models' disagreement predicts test error.
problem Estimating test error of deep networks.
method Empirical testing and theoretical analysis of SGD ensembles.
result SGD ensembles' disagreement correlates with test error.
Stagewise training strategy is widely used for learning neural networks, which runs a stochastic algorithm (e.g., SGD) starting with a relatively large step size (aka learning rate) and geometrically decreasing the step size after a number of iterations. It has been observed that the stagewise SGD has much faster conve…
Paper presents a method to reduce prediction variance of DNNs for unknown systems.
problem Uncertainty in DNN predictions due to high variance.
method Ensemble averaging of multiple DNN models trained independently.
result Reduction in variance of DNN predictions, improving reliability.
Improved sample complexity for training diffusion models.
problem How many samples are needed to train an accurate diffusion model?
method Analyzing the sample complexity of training diffusion models using neural networks.
result Exponential improvement in the dependence on Wasserstein error and depth, along with improved dependencies on other parameters.
Efficient classifier error estimation without re-training.
problem Estimating classifier error without re-training.
method Generalized resubstitution based on empirical measures.
result Consistent and asymptotically unbiased error estimation.
Proposes QEP to mitigate quantization error propagation in layer-wise post-training quantization.
problem Growth of quantization errors across layers degrades performance, especially in low-bit regimes.
method Quantization Error Propagation (QEP) framework that explicitly propagates and compensates for quantization errors.
result QEP-enhanced layer-wise PTQ achieves substantially higher accuracy, especially in low-bit regimes.
Communication is a key bottleneck in distributed training. Recently, an \emph{error-compensated} compression technology was particularly designed for the \emph{centralized} learning and receives huge successes, by showing significant advantages over state-of-the-art compression based methods in saving the communication…
We explore in some detail the notion of algorithmic stability as a viable framework for analyzing the generalization error of learning algorithms. We introduce the new notion of training stability of a learning algorithm and show that, in a general setting, it is sufficient for good bounds on generalization error. In t…
New generalization concept considers distribution of errors, not just average error.
problem Classical generalization fails to capture distributional differences in classifier outputs.
method Formal conjectures about distributional generalization based on model architecture, training procedure, and data distribution.
result Distributional generalization can be expected in specific conditions, as evidenced by empirical results.
Overparameterized models can worsen minority group errors even when overall test error improves.
problem Overparameterization exacerbates spurious correlations, harming minority groups.
method Simulations and experiments on image datasets, theoretical analysis of linear models.
result Subsampling the majority group can achieve low minority error in overparameterized models.
Recent theoretical work has guaranteed that overparameterized networks trained by gradient descent achieve arbitrarily low training error, and sometimes even low test error. The required width, however, is always polynomial in at least one of the sample size n, the (inverse) target error 1/ε, and the (inverse) fail…
SGD-trained deep nets have bounds on their generalization error.
problem Bounding generalization error for deep neural networks trained by SGD.
method Combining dynamical control of parameter norms and Rademacher complexity estimates.
result Explicit bounds depend on loss trajectory, work for various architectures.
Neural network accuracy improves with denser training samples.
problem Improving neural network accuracy on unseen test samples.
method Bounding empirical training error smoothed across activation regions and using it to discard high-risk test samples.
result Discarding high-risk test samples based on error bounds improves prediction accuracy by up to 20%.
Generative AI decodes quantum codes without labeled data.
problem Efficient decoding of quantum error-correcting codes.
method Generative Transformers learn logical operators from unsupervised syndromes.
result Significantly better decoding accuracy than traditional methods.
Sequence-to-sequence models, such as attention-based models in automatic speech recognition (ASR), are typically trained to optimize the cross-entropy criterion which corresponds to improving the log-likelihood of the data. However, system performance is usually measured in terms of word error rate (WER), not log-likel…
This paper introduces a new method for model selection and more generally hyperparameter selection in machine learning. Minimum description length (MDL) is an established method for model selection, which is however not directly aimed at minimizing generalization error, which is often the primary goal in machine learni…
Artificial neural network training with stochastic gradient descent can be destabilized by "bad batches" with high losses. This is often problematic for training with small batch sizes, high order loss functions or unstably high learning rates. To stabilize learning, we have developed adaptive learning rate clipping (A…
This paper tackles worst-class error rate in classification tasks.
problem Minimizing worst-class error rate in classification tasks, especially in medical image classification.
method Designing a boosting approach to bound the worst-class error rate using Deep Neural Networks (DNNs).
result The proposed boosting approach lowers worst-class test error rates while avoiding overfitting.
Data-driven models analyze power grids under incomplete physical information, and their accuracy has been mostly validated empirically using certain training and testing datasets. This paper explores error bounds for data-driven models under all possible training and testing scenarios, and proposes an evaluation implem…
CoNNTrA trains DNNs with low-power, low-memory constraints.
problem Training deep neural networks on edge computing systems with low power and memory usage.
method Coordinate gradient descent-based approach for training DNNs with constrained learning parameters.
result CoNNTrA models use 32x less memory and have comparable errors to Backpropagation models.
Stochastic gradient descent updates parameters with summation gradient computed from a random data batch. This summation will lead to unbalanced training process if the data we obtained is unbalanced. To address this issue, this paper takes the error variance and error mean both into consideration. The adaptively adjus…
Combines cost-sensitive and Neyman-Pearson paradigms for better binary classification.
problem Asymmetric binary classification problems with unequal error severities.
method Develops TUBE-CS algorithm to bridge cost-sensitive and Neyman-Pearson paradigms.
result High-probability control of population type I error.
In-N-Out improves model robustness to out-of-distribution data.
problem Learning robust models with few in-distribution labeled examples.
method Pre-training with auxiliary information and self-training with pseudolabels.
result In-N-Out outperforms auxiliary inputs or outputs alone on both in-distribution and OOD error.
Randomly sampled interpolators achieve zero generalization error with enough data.
problem Understanding the high generalization ability of machine learning models.
method Algebraic geometry tools to prove zero generalization error for random interpolators.
result Generalization error of randomly sampled interpolators becomes zero once the number of training samples exceeds a geometric threshold.
iTimER learns from reconstruction errors to represent irregularly sampled time series.
problem Learning from irregularly sampled time series with missing data.
method iTimER models reconstruction errors as a proxy for unobserved values, using a mixup strategy and a Wasserstein metric.
result iTimER outperforms state-of-the-art methods in classification, interpolation, and forecasting tasks.
This work reduces computation cost for on-device CNN training.
problem High computation cost during on-device CNN training.
method Self-supervised instance filtering and error map pruning.
result Substantial computation saving without significant accuracy loss.
The paper analyzes CycleGAN's error components for unpaired data generation.
problem Analyzing approximation and estimation errors in CycleGAN for unpaired data.
method Decomposes risk into approximation and estimation errors, analyzing each separately and considering their trade-offs.
result Theoretical insights into CycleGAN's performance through error analysis.