arXiv research

The paper proposes a framework for structured prediction using projection oracles.

Orthogonal projections improve learning accuracy in clinical image segmentation and music classification.

Paper develops methods for non-quadratic loss low-rank matrix recovery.

Study geometric properties of loss functions to understand neural network performance.

Proposes PER loss to regularize neural network activations to normal distribution.

Exploits class similarity for better machine learning models with confidence labels and projective loss functions.

Improved online learning with time-varying constraints for complex domains.

Paper tackles online learning on curved spaces without projections.

Improved algorithm reduces communication rounds for distributed online learning.

Compact learning results across various loss functions.

A bidirectional loss function improves label distribution learning and enhancement.

A new method optimizes projection directions for sliced Wasserstein distances.

A novel dictionary-based approach for predicting functions.

Improves deep learning models by blending gradients from training loss and auxiliary objective.

New method trains neural nets without loss functions.

Benguria and Loss have conjectured that, amongst all smooth closed curves of length $2π$ in the plane, the lowest possible eigenvalue of the operator $L=-Δ+κ^2$ was one. They observed that this value was achieved on a two-parameter family, $\mathcal{O}$, of geometrically distinct ovals containing the round circle and c…

Introduces a differentiable approximation to the zero-one loss.

For semi-supervised techniques to be applied safely in practice we at least want methods to outperform their supervised counterparts. We study this question for classification using the well-known quadratic surrogate loss function. Using a projection of the supervised estimate onto a set of constraints imposed by the u…

New approach combines likelihood and adversarial losses for better precipitation predictions.

New research suggests continual learning should focus on both optimization objective and optimization trajectory.

Optimizes differentially private kernel learning with random projection.

Principal Component Analysis (PCA) is a very successful dimensionality reduction technique, widely used in predictive modeling. A key factor in its widespread use in this domain is the fact that the projection of a dataset onto its first $K$ principal components minimizes the sum of squared errors between the original …

SGD updates align with a low-rank subspace but do not lead to further loss reduction.

Optimal convex loss function improves regression coefficient estimation.

A new method bypasses regularization for disentangled latent variables without tuning.

Paper tackles robust classification and feature selection with a novel primal-dual method.

Paper introduces MultiTargeted testing, improving adversarial testing efficiency.

New algorithm reduces adaptive regret without projections.

Performing signal processing tasks on compressive measurements of data has received great attention in recent years. In this paper, we extend previous work on compressive dictionary learning by showing that more general random projections may be used, including sparse ones. More precisely, we examine compressive K-mean…

GD at EoS edge minimizes logistic loss without monotonic convergence.

Algorithm minimizes loss and constraint violations in online convex optimization with smooth penalties.

Optimal weight windows are found by projecting the origin onto a convex polytope.

A new algorithm estimates sparse gradients on graphs with improved risk bounds.

A scalable method for deep metric learning using chance constraints.

New algorithm uses random projections for robust, sparse data classification.

TapNet uses neural networks with task-adaptive projection for improved few-shot learning.

A faster Wasserstein k-means algorithm for histogram data reduces computation and maintains clustering quality.

A new pruning method reduces neural network computation without retraining.

TristouNet is a neural network architecture based on Long Short-Term Memory recurrent networks, meant to project speech sequences into a fixed-dimensional euclidean space. Thanks to the triplet loss paradigm used for training, the resulting sequence embeddings can be compared directly with the euclidean distance, for s…

TTRP method preserves distances in high-dimensional data with reduced storage and speed.

New bounds show linear predictors rarely overfit with certain optimization methods.

Develop a framework for barycentric projections of optimal transport plans on Riemannian manifolds.

This study addresses the challenges of dynamic mini-batch sub-sampling in neural network training.

A new method learns proper multiclass losses and probabilities.

It is now known that an extended Gaussian process model equipped with rescaling can adapt to different smoothness levels of a function valued parameter in many nonparametric Bayesian analyses, offering a posterior convergence rate that is optimal (up to logarithmic factors) for the smoothness class the true function be…

Gradient descent at edge of stability stabilizes implicitly, following projected gradient descent.

We present new excess risk bounds for general unbounded loss functions including log loss and squared loss, where the distribution of the losses may be heavy-tailed. The bounds hold for general estimators, but they are optimized when applied to $η$-generalized Bayesian, MDL, and empirical risk minimization estimators. …

PCA outperforms random projections in retaining second order signals from latent groups.