Unhinged loss minimization fails to improve classifier accuracy for simple data.
problem Accuracy of classifiers minimizing the unhinged loss.
method Minimizing the unhinged loss function.
result Minimizing the unhinged loss yields classifiers with accuracy no better than random guessing for simple data.
Paper analyzes trade-offs in top-k classification accuracies and proposes a new loss function.
problem CE loss does not always optimize top-k prediction, especially with complex data.
method Introduces a novel top-k transition loss to improve top-k accuracy.
result Our loss function improves top-k accuracy, especially for k > 10.
This paper improves forecast stability without sacrificing accuracy using dynamic loss weighting.
problem Rolling origin forecast instability in time series forecasting.
method Dynamic loss weighting algorithms applied to the N-BEATS model.
result Dynamic loss weighting can further improve forecast stability without compromising accuracy.
New expressive losses improve adversarial robustness without sacrificing accuracy.
problem Training networks for robustness at the expense of accuracy.
method Formalizing expressivity, using convex combinations of adversarial attacks and IBP bounds.
result Trivial expressive losses yield state-of-the-art results in various settings.
New loss function improves accuracy of MRI parameter estimation.
problem Systematic errors in parameter estimates at low SNR.
method Developed and implemented negative log Rician likelihood (NLR) loss.
result NLR loss shows higher accuracy in parameter estimation than MSE loss at low SNR.
Proposes squentropy loss for improved classification accuracy and model calibration.
problem Theoretical and empirical evidence for cross-entropy loss is lacking.
method Introduces squentropy loss as the sum of cross-entropy and average square loss over incorrect classes.
result Squentropy loss outperforms cross-entropy and rescaled square losses in classification accuracy and model calibration.
Focal loss improves deep neural networks' accuracy and calibration.
problem Miscalibration in deep neural networks.
method Using focal loss and temperature scaling to improve model calibration.
result Focal loss leads to state-of-the-art calibrated models without sacrificing accuracy.
PACMAN provides bounds for classification tasks considering accuracy vs. negative log-loss mismatch.
problem Mismatch between accuracy and negative log-loss in classification tasks.
method Point-wise PAC approach over generalization gap, using likelihood ratio and concentration inequalities.
result PACMAN provides point-wise PAC bounds for the generalization problem.
SGD with large learning rates can achieve better test accuracy than expected.
problem SGD with large learning rates often outperforms expected convergence bounds.
method Proved that SGD with small learning rates stays close to gradient flow path on modified loss.
result Explicitly adding an implicit regularizer to the loss improves test accuracy.
Proposes new loss functions for GANs to improve estimation accuracy and robustness.
problem Improving the training of GANs to achieve more accurate and robust models.
method Introduces Hellinger-type loss functions and analyzes their statistical properties.
result Demonstrates improved estimation accuracy and robustness of the proposed loss functions.
The 01 loss is robust to outliers and tolerant to noisy data compared to convex loss functions. We conjecture that the 01 loss may also be more robust to adversarial attacks. To study this empirically we have developed a stochastic coordinate descent algorithm for a linear 01 loss classifier and a single hidden layer 0…
Out-of-distribution (OOD) detection approaches usually present special requirements (e.g., hyperparameter validation, collection of outlier data) and produce side effects (e.g., classification accuracy drop, slower energy-inefficient inferences). We argue that these issues are a consequence of the SoftMax loss anisotro…
Proposes a hierarchical curriculum loss to improve model accuracy and interpretability.
problem Flat label spaces in classification algorithms fail to capture dependencies in real-world data.
method Introduces hierarchical curriculum loss with two properties: satisfying hierarchical constraints and providing non-uniform label weights.
result The proposed loss function significantly outperforms multiple baselines on real-world image datasets.
New insights into convergence and accuracy trade-offs in federated and meta-learning.
problem Understanding the trade-offs between convergence and accuracy in federated and meta-learning.
method Generalized local update methods, proving equivalence to first-order optimization on a surrogate loss.
result Novel convergence rates and insights into the importance of algorithmic choices in communication-limited settings.
SAT improves adversarial training by smoothing the loss landscape through curriculum learning.
problem Adversarial training sacrifices clean accuracy for robustness and suffers from large generalization error.
method SAT uses curriculum learning to smooth the adversarial loss landscape, improving both clean and robust accuracy.
result SAT models improve clean and robust accuracy significantly compared to adversarial training and other baselines.
Custom loss functions improve accuracy of wildfire rate of spread forecasts.
problem Improving accuracy of wildfire rate of spread forecasts.
method Examined custom loss functions in machine learning models of fuel moisture content.
result Custom loss functions improved accuracy of ROS forecasts by a small amount.
Enhances ENet's prediction accuracy while maintaining uncertainty estimation.
problem Gradient shrinkage problem in ENet's loss function.
method Proposes a multi-task learning framework with a modified MSE loss function.
result Improves ENet's prediction accuracy without losing uncertainty estimation.
This paper analyzes the conflict between Hamming loss and subset accuracy in multi-label classification.
problem The conflict between Hamming loss and subset accuracy in multi-label classification.
method The paper analyzes the learning guarantees of algorithms optimizing Hamming loss and subset accuracy, providing theoretical bounds and experimental support.
result Optimizing Hamming loss with its surrogate loss can lead to good performance on subset accuracy in small label spaces, contrary to theoretical expectations.
New methods connect low-loss points on neural network surfaces.
problem Connecting low-loss points on neural network loss surfaces.
method Macroscopic distributional assumptions and global connection models.
result Accuracy correlates with complexity and sensitivity.
Paper introduces negative margin loss for better few-shot classification accuracy.
problem Improving few-shot classification accuracy with metric learning.
method Introduces negative margin loss and analyzes its impact on feature discriminability.
result Negative margin loss outperforms regular softmax loss on few-shot classification benchmarks.
Convolution Neural Networks (CNN) have recently achieved state-of-the art performance on handwritten Chinese character recognition (HCCR). However, most of CNN models employ the SoftMax activation function and minimize cross entropy loss, which may cause loss of inter-class information. To cope with this problem, we pr…
In classification, the de facto method for aggregating individual losses is the average loss. When the actual metric of interest is 0-1 loss, it is common to minimize the average surrogate loss for some well-behaved (e.g. convex) surrogate. Recently, several other aggregate losses such as the maximal loss and average t…
Supervised contrastive learning improves image classification accuracy.
problem Improving image classification accuracy using supervised contrastive learning.
method Extending self-supervised batch contrastive approach to fully-supervised setting, leveraging label information.
result Top-1 accuracy of 81.4% on ImageNet dataset, outperforming cross-entropy.
Study reveals how initialization scale affects training accuracy in linear networks.
problem Understanding implicit bias in linear classification models.
method Asymptotic analysis of gradient flow trajectories and training loss minimization.
result Implicit bias is more complex at reasonable initialization scales and training accuracies.
New loss improves DNN calibration without sacrificing accuracy.
problem Overfitting and overconfident predictions in DNNs.
method Proposes a new loss function inspired by Bayes decision theory.
result Improves DNN calibration without compromising accuracy.
We develop a novel method, called PoWER-BERT, for improving the inference time of the popular BERT model, while maintaining the accuracy. It works by: a) exploiting redundancy pertaining to word-vectors (intermediate encoder outputs) and eliminating the redundant vectors. b) determining which word-vectors to eliminate …
Multi-output is essential in machine learning that it might suffer from nonconforming residual distributions, i.e., the multi-output residual distributions are not conforming to the expected distribution. In this paper, we propose "Wrapped Loss Function" to wrap the original loss function to alleviate the problem. This…
New binary loss functions improve density ratio estimation accuracy.
problem Improving accuracy of density ratio estimators using binary classifiers.
method Characterized loss functions based on prescribed error measures in Bregman divergences.
result Novel loss functions prioritize accurate estimation of large density ratio values.
The loss functions of deep neural networks are complex and their geometric properties are not well understood. We show that the optima of these complex loss functions are in fact connected by simple curves over which training and test accuracy are nearly constant. We introduce a training procedure to discover these hig…
The Gradient Boosting Decision Tree (GBDT) is a popular machine learning model for various tasks in recent years. In this paper, we study how to improve model accuracy of GBDT while preserving the strong guarantee of differential privacy. Sensitivity and privacy budget are two key design aspects for the effectiveness o…
A new loss function improves neural networks' out-of-distribution detection without side effects.
problem Neural networks struggle with out-of-distribution detection due to SoftMax loss issues.
method Proposes IsoMax loss replacing SoftMax loss, maintaining high entropy and fast inferences.
result Significantly improves neural networks' out-of-distribution detection performance.
We analyze the fluctuation of the loss from default around its large portfolio limit in a class of reduced-form models of correlated firm-by-firm default timing. We prove a weak convergence result for the fluctuation process and use it for developing a conditionally Gaussian approximation to the loss distribution. Nume…
We propose a novel approach for loss reserving based on deep neural networks. The approach allows for joint modeling of paid losses and claims outstanding, and incorporation of heterogeneous inputs. We validate the models on loss reserving data across lines of business, and show that they improve on the predictive accu…
In this paper, we propose a novel progressive parameter pruning method for Convolutional Neural Network acceleration, named Structured Probabilistic Pruning (SPP), which effectively prunes weights of convolutional layers in a probabilistic manner. Unlike existing deterministic pruning approaches, where unimportant weig…
Previous work shows that adversarially robust generalization requires larger sample complexity, and the same dataset, e.g., CIFAR-10, which enables good standard accuracy may not suffice to train robust models. Since collecting new training data could be costly, we focus on better utilizing the given data by inducing t…
In this paper, we propose a novel {\em e-exponentiated} transformation, 0≤e<1, for loss functions. When the transformation is applied to a convex loss function, the transformed loss function become more robust to outliers. Using a novel generalization error bound, we have theoretically shown that the transforme…
We propose a generic framework to calibrate accuracy and confidence of a prediction in deep neural networks through stochastic inferences. We interpret stochastic regularization using a Bayesian model, and analyze the relation between predictive uncertainty of networks and variance of the prediction scores obtained by …
The article provides formulas to hedge impermanent loss in decentralized markets.
problem Impermanent loss in concentrated liquidity provision in decentralized markets.
method Analytical characterizations and static replication formulas using European calls or puts.
result Static replication formulas accurately hedge impermanent loss.
A new asymmetric contrastive loss improves performance on imbalanced datasets.
problem Improving performance on imbalanced datasets using contrastive learning.
method Introducing an asymmetric contrastive loss (ACL) and asymmetric focal contrastive loss (AFCL).
result AFCL outperforms CL and FCL in terms of weighted and unweighted classification accuracies on imbalanced datasets.
In this paper we propose a method of obtaining points of extreme overfitting - parameters of modern neural networks, at which they demonstrate close to 100 % training accuracy, simultaneously with almost zero accuracy on the test sample. Despite the widespread opinion that the overwhelming majority of critical points o…
Improved robustness of machine learning models with controlled Lipschitz constants.
problem Vulnerability of state-of-the-art models to adversarial attacks.
method Proposes a CLL loss that calibrates the margin and Lipschitz constant penalties, improving robustness certificates.
result Consistently outperforms other losses on CIFAR-10, CIFAR-100, and Tiny-ImageNet datasets.
Deep models can fit noisy labels, but robustness and reliability are still issues.
problem Training deep models with noisy labels leads to unreliable uncertainty quantification.
method Analysis of conditional distribution over noisy labels and evaluation of robust loss functions.
result Strictly proper and robust loss functions preserve accuracy but do not guarantee reliability.
01 loss neural network improves adversarial robustness on CIFAR10.
problem Improving machine learning models' resistance to adversarial attacks.
method Proposes a single hidden layer 01 loss neural network trained with stochastic coordinate descent.
result 01 loss network has the largest adversarial distortion compared to other models.
Recurrent Neural Networks (RNN) can be difficult to deploy on resource constrained devices due to their size.As a result, there is a need for compression techniques that can significantly compress RNNs without negatively impacting task accuracy. This paper introduces a method to compress RNNs for resource constrained e…
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.
New loss function restores importance weighting in overparameterized models.
problem Restoring importance weighting in overparameterized neural networks.
method Introduced polynomially-tailed losses to restore effects of importance weighting.
result Polynomially-tailed losses improve performance in correcting distribution shift.
Enhances deep learning robustness to noise without sacrificing clean data accuracy.
problem Robustness of deep neural networks to input noise.
method Discriminative loss at penultimate layer and class-wise feature alignment with Gaussian noise.
result Improves robustness to various perturbations without degrading clean data accuracy.
SoftAD improves classification accuracy with less fine-tuning and fewer computational costs.
problem Improving classification accuracy with less fine-tuning and fewer computational costs.
method SoftAD is a softened, pointwise mechanism that downweights borderline points and limits the effects of outliers.
result SoftAD achieves classification accuracy competitive with flooding and SAM, with a smaller loss generalization gap and model norm.