Least squares regression shows unexpected double descent in under-parameterized models.
problem Understanding the generalization of under-parameterized models in regression.
method Analyzing the spectrum and eigenvectors of the sample covariance matrix.
result Least squares regression can exhibit a peak in generalization in the under-parameterized regime, contrary to previous explanations.
Under-parameterization hinders deep RL's efficiency.
problem Implicit under-parameterization impairs data-efficiency in deep RL.
method Characterized and mitigated the rank collapse of value network features.
result Controlling rank collapse improves deep RL performance.
New algorithms improve binary neural network configurations.
problem Training binary neural networks efficiently.
method Stochastic message passing algorithms (BP and SP) for discrete inference.
result Stochastic BP and SP find better BNN configurations.
Small parameterized towers improve multi-task learning efficiency and generalization.
problem Balancing Pareto efficiency and generalization in multi-task learning.
method Under-parameterized self-auxiliaries for multi-task models.
result Small parameterized towers enhance Pareto efficiency in various multi-task applications.
Under-parameterized networks can either copy or average teacher weights, leading to universal optimal solutions.
problem Approximating a teacher network with an under-parameterized student network.
method Analyzing shallow neural networks with erf activation function and unitary teacher weights, proving copy-average configurations are critical points and finding the optimal solution.
result The optimal solution for under-parameterized networks has a universal structure, whether copying or averaging teacher neurons.
New insights into optimization and generalization for linear models.
problem Understanding the implicit regularization of optimization methods for linear models.
method Investigating the norms minimized by interpolating solutions and using projections to move between solutions.
result Proving that for over-parameterized linear classification, projections onto the data-span enable the use of under-parameterized techniques.
Improved neural network training by coupled initialization reduces neuron count.
problem Training neural networks efficiently with fewer neurons.
method Coupled initialization of weights into pairs of identical Gaussian vectors.
result Significantly reduced number of neurons required for network convergence.
The paper shows how the generalization curve can have multiple peaks, influenced by data and learning algorithm biases.
problem Understanding the generalization behavior of linear regression models under varying parameterizations.
method Analyzes generalization loss in linear regression models with varying parameterizations, both under- and over-parameterized.
result The generalization curve can have an arbitrary number of peaks, and their locations can be controlled.
Although deep learning has shown its powerful performance in many applications, the mathematical principles behind neural networks are still mysterious. In this paper, we consider the problem of learning a one-hidden-layer neural network with quadratic activations. We focus on the under-parameterized regime where the n…
COCOA can learn new tasks without forgetting previously learned ones in a distributed setting.
problem Continual learning with tasks arriving sequentially.
method Distributed optimization with COCOA algorithm.
result COCOA can perform continual learning without forgetting.
GD learns matrix solutions incrementally, revealing insights into generalization.
problem Matrix sensing problem of recovering low-rank matrices from linear measurements.
method Fine-grained analysis of GD dynamics for matrix sensing.
result GD follows an incremental learning procedure, solving matrices of increasing ranks.
New method calibrates neural network predictions for better reliability.
problem Improper probability estimates from deep networks leading to unreliable predictions.
method Proposes a constrained optimization approach for a monotonic calibration map.
result Achieves state-of-the-art performance across various datasets and models.
Deep networks can handle noisy labels up to a certain threshold.
problem Deep learning's robustness to noisy labels.
method Applying classical statistical theory and universal consistency of DNNs.
result Certain DNNs can tolerate massive symmetric label noise up to the information-theoretic threshold.
New framework explains diverse impacts of data augmentation.
problem Understanding the varied effects of data augmentation on model performance.
method Developed a theoretical framework to characterize DA's impact on linear models.
result Data augmentation induces implicit spectral regularization through two effects.
New bounds for transfer learning in linear models, improving generalization.
problem Understanding when auxiliary data helps in improving generalization in linear models.
method Derivation of exact error bounds and optimal task weights for linear regression and linear neural networks.
result First non-vacuous sufficient conditions for beneficial auxiliary learning in linear neural networks.
The study uncovers the conditions under which diffusion models memorize or generalize.
problem Understanding the balance between memorization and generalization in diffusion models.
method Theoretical and mathematical framework to investigate memorization and generalization in diffusion models.
result Theoretical crossover point predicts a phase transition in diffusion models, validating the hypothesis.
Linear models can overfit without harming OOD generalization under certain conditions.
problem Understanding how overparameterized linear models generalize to out-of-distribution data.
method Analyzing overparameterized linear models under covariate shift, providing guarantees for OOD generalization.
result Benign overfitting occurs in standard ridge regression under OOD conditions, with specific structural conditions on target covariance.
This paper extends stability and generalization analysis of GD for multi-layer NNs.
problem Understanding the generalization of multi-layer neural networks trained by GD.
method Comprehensive stability and generalization analysis of GD for multi-layer NNs, focusing on two-layer and three-layer networks.
result Derives excess risk rates of O ( 1 / n ) O(1/\sqrt{n}) O ( 1/ n ) for GD in two-layer and three-layer NNs under specific conditions. Paper analyzes gradient descent with noisy data copies for linear regression, showing regularization and acceleration effects.
problem Improving generalization in machine learning through data augmentation with noise.
method Gradient descent with on-line noisy copies for linear regression analysis.
result Training with on-line noisy copies is equivalent to ridge regularization with a specific regularization parameter.