Enhances linear regression with Kalman filter for loss minimization.
problem Minimizing loss in linear regression models.
method Integrates Kalman filter and SGD for optimal weight updates.
result Develops optimal linear regression equation with minimum area under curve.
Calibrating classifiers reduces grouping loss using sufficiency criteria.
problem Grouping loss in probabilistic classifier calibration is often overlooked.
method Revisited Langford & Zadrozny's probing reduction approach and introduced Brier curves.
result The probing reduction approach reduces grouping loss and supports sufficient calibration.
We propose a max-pooling based loss function for training Long Short-Term Memory (LSTM) networks for small-footprint keyword spotting (KWS), with low CPU, memory, and latency requirements. The max-pooling loss training can be further guided by initializing with a cross-entropy loss trained network. A posterior smoothin…
Analyzes double descent in binary classification models with different losses.
problem Understanding the double descent phenomenon in binary classification models.
method Analytic study of gradient descent with logistic and square losses on binary linear classification models.
result The double descent phenomenon persists but with differences compared to logistic loss.
Flooding prevents deep networks from achieving zero training loss, improving test performance.
problem Deep networks can achieve zero training error but often have zero training loss, leading to overconfidence and poor test performance.
method Flooding prevents training loss from reaching zero by applying gradient ascent when loss is below a preset flood level.
result Flooding improves test performance and induces a double descent curve of the test loss.
In this manuscript we propose two objective terms for neural image compression: a compression objective and a cycle loss. These terms are applied on the encoder output of an autoencoder and are used in combination with reconstruction losses. The compression objective encourages sparsity and low entropy in the activatio…
The paper predicts loss scaling across different datasets and compute scales.
problem Predicting loss scaling across different datasets and compute scales.
method Derive shifted power law relationships between train and test losses.
result Shifted power law relationships hold for various datasets and tasks, improving prediction accuracy.
This paper identifies and analyzes the Epochal Sawtooth Phenomenon in training loss curves.
problem Training loss oscillations in adaptive gradient-based optimizers.
method Empirical analysis of Adam and other optimizers, focusing on β parameters, batch size, data shuffling, and sample replacement. result The Epochal Sawtooth Phenomenon (ESP) is a recurring pattern in training loss curves, arising from adaptive learning rate adjustments and data shuffling.
The paper presents a multi-power law for predicting loss curves across different learning rate schedules.
problem Understanding and optimizing the relationship between model performance and hyperparameters, especially learning rates.
method Proposes a multi-power law that combines power laws based on the sum of learning rates and additional laws for loss reduction due to decay.
result The multi-power law accurately predicts loss curves for unseen learning rate schedules and finds a schedule that outperforms cosine learning rate.
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…
Model shows loss curve with two distinct exponents due to sparse activations.
problem Sparse activations impact neural network scaling laws.
method Introduced a model for neural scaling laws under sparse activations, derived asymptotic population loss, and analyzed gradient-descent dynamics.
result Loss curve exhibits double-descent peak near interpolation threshold with two distinct scaling exponents.
Geodesics connect model modes in neural network loss landscapes.
problem Connecting modes in neural network loss landscapes.
method Reframed mode connectivity in Information Geometry, hypothesized geodesics as mode-connecting paths, proposed algorithm to approximate geodesics.
result Geodesics achieve mode connectivity in neural networks.
New model predicts neural network performance from early training epochs, incorporating architecture impact.
problem Predicting neural network performance from early training epochs, neglecting architecture impact.
method Architecture-aware graph ordinary differential equation model.
result Model outperforms state-of-the-art methods for MLP and CNN learning curves.
Symmetric losses improve classifier robustness from corrupted labels.
problem Improving classifier performance from corrupted labels.
method Symmetric losses that satisfy a certain condition.
result Symmetric losses enhance robust classification from corrupted labels.
Modified AUC improves CNN training by considering model confidence.
problem Improving binary classifier performance metrics.
method Proposes a modified AUC metric that incorporates model confidence into BCE loss for CNN training.
result Demonstrates improved performance on three datasets: MNIST, prostate MRI, and brain MRI.
Study predicts SGD test loss for structured features.
problem Understanding test loss dynamics in SGD for structured data.
method Solveable model of SGD on mean square loss for arbitrary covariance structure.
result Simpler Gaussian model accurately predicts test loss of nonlinear models.
This work reveals how label noise can cause a final ascent in neural network performance curves.
problem The impact of label noise on the performance of neural networks.
method Theoretical analysis and extensive experiments on various neural network architectures.
result Label noise can lead to a final ascent in the test loss curve, improving generalization at intermediate model widths.
The study examines the dynamic behavior of RMSprop and Adam algorithms.
problem Understanding the training loss curve patterns of adaptive gradient algorithms.
method Careful numerical experiments and theoretical explanations using the signGD flow.
result Adam converges smoother and faster when momentum factors are close to each other.
New method saves computational budget by ranking and transferring learning curves.
problem Expensive automated machine learning methods for hyperparameter and neural architecture optimization.
method Tackles as a ranking and transfer learning problem, optimizing a pairwise ranking loss and leveraging learning curves from other datasets.
result Accelerates neural architecture search by a factor of up to 100 without significant performance degradation.
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.
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.
Mode connectivity is a recently introduced frame- work that empirically establishes the connected- ness of minima by finding a high accuracy curve between two independently trained models. To investigate the limits of this setup, we examine the efficacy of this technique in extreme cases where the input models are trai…
We present a novel approach to adjust global image properties such as colour, saturation, and luminance using human-interpretable image enhancement curves, inspired by the Photoshop curves tool. Our method, dubbed neural CURve Layers (CURL), is designed as a multi-colour space neural retouching block trained jointly in…
Deep networks can interpolate noisy data without losing generalization.
problem Characterizing the relationship between interpolation and generalization in overparameterized deep networks.
method Analyzing the loss landscape of neural network functions over volumes around training data points, varying model parameters and training epochs.
result Loss sharpness in the input space follows a double descent, with large models predicting noisy targets over larger volumes around training data points.
New models improve classification model performance, especially robust to small training sets.
problem Improving classification model performance, especially robust to small training sets.
method Distributionally robust AUC maximization models using Kantorovich metric and hinge loss function.
result The proposed DR-AUC models outperform standard models in general and worst-case out-of-sample performance.
LoRA-Curve connects independent LoRA optima through continuous low-loss valleys, improving Bayesian model averaging.
problem Challenges in estimating epistemic uncertainty in LoRA-based Bayesian inference.
method Introduces LoRA-Curve, a segmented Bézier curve parameterization in the LoRA space, with free and anchored configurations.
result Empirically shows that connecting independent LoRA optima through continuous low-loss valleys improves mutual information of the predictive distribution.
We analyze impermanent loss in AMMs and show G3Ms are simplest.
problem Understanding impermanent loss in automated market makers.
method Developed a general framework and analyzed Geometric Mean Market Makers (G3Ms).
result G3Ms have the simplest impermanent loss characteristics.
Transformers exhibit abrupt learning in matrix completion tasks.
problem Understanding abrupt learning in Transformers for matrix completion.
method Formulated matrix completion as MLM task, trained BERT model, analyzed model components.
result Sudden drop in loss despite no changes in training procedure or hyper-parameters.
Study how data structure impacts classification performance in overparametrized models.
problem Understanding the role of data structure in overparametrized models.
method Analytical and statistical physics methods applied to structured data.
result Data structure significantly affects the performance of overparametrized models, particularly for logistic loss.
Investment horizon approach has been used to analyze indexes of Polish stock market.Optimal time horizon for each return value is evaluated by fitting appropriate function form of the distribution. Strong asymmetry of gain-loss curves is observed for WIG index, whereas gain and loss curves look similar for WIG20 and fo…
New insights into deep learning: reducing training data significantly improves performance.
problem Understanding and improving generalization in deep learning models.
method Analyzing the distribution of classification margins and dynamically reducing the training set.
result The area under the curve of the margin distribution is a good measure of generalization.
Comparison of decision curve analysis and cost curves for model evaluation.
problem Evaluating classification performance across different operating contexts.
method Comparison of Decision Curve Analysis (DCA) and Cost Curves.
result DCA and Cost Curves are closely related, with Brier curves being more generally applicable.
Proposes neuron alignment to optimize mode connectivity in neural networks.
problem Understanding and optimizing mode connectivity in deep neural networks.
method Introduces neuron alignment to approximate optimal weight permutations and improve mode connectivity.
result Neuron alignment significantly alleviates robust loss barriers and improves model robustness and accuracy.
AGF explains feature learning in neural networks through alternating steps.
problem Understanding what features neural networks learn and how they learn them.
method AGF is an algorithmic framework that approximates the dynamics of feature learning in two-layer networks.
result AGF provides a unified framework to understand feature learning in neural networks, matching experimental results across various architectures.
New insights into hidden minima in neural networks.
problem Identifying hidden minima in two-layer ReLU networks.
method Analyzing curves along which loss is minimized, focusing on eigenvalue contributions.
result Distinctive structural and symmetry properties of arcs emanating from hidden minima.
New optimization method improves AUC for binary classification and changepoint detection.
problem Non-convex AUC and sub-optimal points in ROC curves.
method AUM (Area Under Min(FP, FN)) surrogate loss function based on sorting and summing ROC curve points.
result AUM minimization learning algorithm improves AUC and speeds up compared to previous methods.
Bayesian method corrects bias in imbalanced datasets.
problem Prevalence bias in machine learning datasets.
method Bayesian risk minimization framework, bias-corrected loss function.
result Corrected loss function improves model performance.
New method improves fairness of facial recognition systems.
problem Facial recognition systems exhibit bias across different demographic groups.
method Optimizes centroid-based scores to reduce bias in pre-trained models.
result Demonstrates significant improvement in fairness with minimal loss in accuracy.
More training data can hurt the generalization of adversarially robust models.
problem The challenge of balancing adversarial robustness and generalization in machine learning models.
method Investigation of three regimes based on adversary strength and empirical studies on various models.
result More training data can hurt the generalization of adversarially robust models in different regimes.
Paper tackles temporal overfitting in wind power curve modeling.
problem Temporal overfitting in wind power curve modeling.
method Proposes a Gaussian process-based method to partition and model time-invariant and time-varying components.
result Significant improvement in predicting responses for different time periods.
New methods evaluate data representations by complexity of low-loss predictor learning.
problem Evaluating quality of data representations for downstream tasks.
method Surplus Description Length (SDL) and ε Sample Complexity (εSC) methods.
result Methods measure the information needed to approximate optimal predictor up to specified tolerance.
New approach uses hinge loss for iterative regularization in classification.
problem Improving classification accuracy through regularization.
method Develops an iterative regularization approach based on hinge loss.
result Proves convergence and rates of convergence for classification.
Adaptive market maker curves minimize arbitrage losses in DeFi.
problem Asset trading prices in AMMs trail behind centralized exchanges, causing LP losses.
method Adapts market maker bonding curves to trader behavior using a differential equation derived from the Glosten-Milgrom model.
result Optimal adaptive curves minimize arbitrage losses while remaining competitive.
Burq-Gérard-Tzvetkov and Hu established Lp estimates (2≤p≤∞) for the restriction of eigenfunctions to submanifolds. The estimates are sharp, except for the log loss at the endpoint L2 estimates for submanifolds of codimension 2. It has long been believed that the log loss at the endpoint can be remov…
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+κ2 was one. They observed that this value was achieved on a two-parameter family, O, of geometrically distinct ovals containing the round circle and c…
New method optimizes lossy compression models more effectively.
problem Difficulty in manually setting a tradeoff parameter for lossy compression models.
method Constrained optimization method of Rezende and Viola, 2018.
result The method enables pointwise model comparisons and outperforms existing methods.
We provide a general framework for characterizing the trade-off between accuracy and robustness in supervised learning. We propose a method and define quantities to characterize the trade-off between accuracy and robustness for a given architecture, and provide theoretical insight into the trade-off. Specifically we in…
In this work, a dense recurrent convolutional neural network (DRCNN) was constructed to detect sleep disorders including arousal, apnea and hypopnea using Polysomnography (PSG) measurement channels provided in the 2018 Physionet challenge database. Our model structure is composed of multiple dense convolutional units (…