arXiv research

DLM for BNNs fails to improve over ELBO optimization.

Gradient descent implicitly follows regularization for general losses.

The paper improves sparse Gaussian processes by optimizing predictive loss.

This work improves structured prediction by learning the balance between signal and random noise.

GD at EoS edge minimizes logistic loss without monotonic convergence.

Reparameterization of variational auto-encoders with continuous random variables is an effective method for reducing the variance of their gradient estimates. In the discrete case, one can perform reparametrization using the Gumbel-Max trick, but the resulting objective relies on an $\arg \max$ operation and is non-dif…

Boosted CVaR Classification improves tail performance in classification tasks.

A new pruning method finds sparse minimizers in flat regions of deep neural networks.

Training activation quantized neural networks involves minimizing a piecewise constant function whose gradient vanishes almost everywhere, which is undesirable for the standard back-propagation or chain rule. An empirical way around this issue is to use a straight-through estimator (STE) (Bengio et al., 2013) in the ba…

The past decade has witnessed a successful application of deep learning to solving many challenging problems in machine learning and artificial intelligence. However, the loss functions of deep neural networks (especially nonlinear networks) are still far from being well understood from a theoretical aspect. In this pa…

Task loss matching misrepresents similarity between neural network layers.

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

A new method optimizes neural sequence models for better task performance.

Stochastic Gradient Descent (SGD) based training of neural networks with a large learning rate or a small batch-size typically ends in well-generalizing, flat regions of the weight space, as indicated by small eigenvalues of the Hessian of the training loss. However, the curvature along the SGD trajectory is poorly und…

Symmetric losses improve classifier robustness from corrupted labels.

Deep generative networks can simulate from a complex target distribution, by minimizing a loss with respect to samples from that distribution. However, often we do not have direct access to our target distribution - our data may be subject to sample selection bias, or may be from a different but related distribution. W…

Autoencoders are a deep learning model for representation learning. When trained to minimize the distance between the data and its reconstruction, linear autoencoders (LAEs) learn the subspace spanned by the top principal directions but cannot learn the principal directions themselves. In this paper, we prove that $L_2…

A new pruning method reduces neural network computation without retraining.

TSSM splits neural networks for parallel training with minimal accuracy loss.

Transfer learning have been frequently used to improve deep neural network training through incorporating weights of pre-trained networks as the starting-point of optimization for regularization. While deep transfer learning can usually boost the performance with better accuracy and faster convergence, transferring wei…

We correct for sampling bias in training models to improve real-world performance.

Proposes a method to prevent overfitting in deep DRE models.

Unified algorithm for tensor decomposition supports multiple loss functions and models.

We propose a risk-averse statistical learning framework wherein the performance of a learning algorithm is evaluated by the conditional value-at-risk (CVaR) of losses rather than the expected loss. We devise algorithms based on stochastic gradient descent for this framework. While existing studies of CVaR optimization …

We propose a new regularization method based on virtual adversarial loss: a new measure of local smoothness of the conditional label distribution given input. Virtual adversarial loss is defined as the robustness of the conditional label distribution around each input data point against local perturbation. Unlike adver…

This work justifies neural collapse under MSE loss and analyzes the optimization landscape.

Deep networks converge in direction, with implications for predictions and margins.

In this paper, we address the problem of embedded feature selection for ranking on top of the list problems. We pose this problem as a regularized empirical risk minimization with $p$-norm push loss function ($p=\infty$) and sparsity inducing regularizers. We leverage the issues related to this challenging optimization…

Improved online algorithm for convex losses with near-optimal swap regret.

We extend the well-known BFGS quasi-Newton method and its memory-limited variant LBFGS to the optimization of nonsmooth convex objectives. This is done in a rigorous fashion by generalizing three components of BFGS to subdifferentials: the local quadratic model, the identification of a descent direction, and the Wolfe …

Develops a direct debiased machine learning framework using Bregman divergence.

A property, or statistical functional, is said to be elicitable if it minimizes expected loss for some loss function. The study of which properties are elicitable sheds light on the capabilities and limitations of point estimation and empirical risk minimization. While recent work asks which properties are elicitable, …

ScoreStop uses gradient tests to stop gradient boosting early.

New PG losses improve decision optimization in misspecified models.

We prove that the empirical risk of most well-known loss functions factors into a linear term aggregating all labels with a term that is label free, and can further be expressed by sums of the loss. This holds true even for non-smooth, non-convex losses and in any RKHS. The first term is a (kernel) mean operator --the …

This paper fortifies the recently introduced hierarchical-optimization recursive least squares (HO-RLS) against outliers which contaminate infrequently linear-regression models. Outliers are modeled as nuisance variables and are estimated together with the linear filter/system variables via a sparsity-inducing (non-)co…

New PAC-Bayes bounds for unbounded losses using Cramér-Chernoff techniques.

Unhinged loss minimization fails to improve classifier accuracy for simple data.

New proof shows D-SGD and SAM are equivalent, revealing advantages of decentralization.

Support vector machine (SVM) has attracted great attentions for the last two decades due to its extensive applications, and thus numerous optimization models have been proposed. To distinguish all of them, in this paper, we introduce a new model equipped with an $L_{0/1}$ soft-margin loss (dubbed as $L_{0/1}$-SVM) whic…

Study reveals sharp characterisation of local minima in neural network loss landscapes.

We propose a stochastic optimization method for minimizing loss functions, expressed as an expected value, that adaptively controls the batch size used in the computation of gradient approximations and the step size used to move along such directions, eliminating the need for the user to tune the learning rate. The pro…

Summarizing large-scaled directed graphs into small-scale representations is a useful but less studied problem setting. Conventional clustering approaches, which based on "Min-Cut"-style criteria, compress both the vertices and edges of the graph into the communities, that lead to a loss of directed edge information. O…

Differentiable structure learning addresses DAGs with multiple global minimizers.

This work addresses unstable MeanFlow training by optimizing a coefficient in the loss function.

New algorithms minimize PAC-Bayesian C-Bound for majority voting, leading to scalable and accurate predictors.

A novel linear classification method that possesses the merits of both the Support Vector Machine (SVM) and the Distance-weighted Discrimination (DWD) is proposed in this article. The proposed Distance-weighted Support Vector Machine method can be viewed as a hybrid of SVM and DWD that finds the classification directio…

In this paper, we propose a novel implicit semantic data augmentation (ISDA) approach to complement traditional augmentation techniques like flipping, translation or rotation. Our work is motivated by the intriguing property that deep networks are surprisingly good at linearizing features, such that certain directions …