Catapult phase in neural nets shows exponential loss growth before quick decrease.
problem Understanding phase transitions in neural networks during training.
method Analyzing weight norm and loss behavior for super-critical learning rates.
result Proven existence of catapult phase in quadratic models and two-layer nets.
Study shows exponential error reduction in multiclass classification without bias-variance trade-off.
problem Multiclass classification with margin conditions.
method Analysis of classification error under hard-margin conditions.
result Exponential decrease in classification error without bias-variance trade-off.
A new topology improves decentralized learning efficiency and accuracy.
problem Finding efficient decentralized learning topologies with fast consensus and low maximum degree.
method Proposed the Base-(k+1) Graph topology for decentralized learning. result The Base-(k+1) Graph enables faster convergence and better communication efficiency than the exponential graph. Paper establishes universal lower bounds and optimal rates for clustering sub-exponential mixture models.
problem Achieving optimal error rates in clustering sub-exponential mixture models.
method Establishes universal lower bounds and demonstrates iterative algorithms' optimality in sub-exponential mixture models.
result Iterative algorithms achieve the universal lower bound in sub-exponential mixture models.
Paper improves deep learning convergence rates for low-dimensional data.
problem Sub-optimal rates in deep learning due to unrealistic assumptions on intrinsic dimension.
method Introduced an entropic notion of intrinsic dimension for exponential families and demonstrated improved convergence rates.
result Test error scales as O~(n−2β+dˉ2β(λ)2β), improving on best-known rates. Paper shows SVM can achieve super fast convergence rates.
problem Understanding fast convergence rates for SVM.
method Presented a simple mechanism to obtain fast convergence rates for SVM.
result SVM can exhibit exponential convergence rates without hard Tsybakov margin condition.
Paper proposes an algorithm to recover full supervision from weakly labeled data.
problem Machine learning requires expensive data annotation, motivating the use of weak supervision.
method The paper introduces a disambiguation principle and an empirical disambiguation algorithm for partial labelling.
result The algorithm achieves exponential convergence rates under learnability assumptions.
New learning rate schedule improves deep learning performance.
problem Improving deep learning performance with varying learning rates.
method Exponential learning rate schedule with Batch Normalization.
result Exponential learning rate schedule with BN is equivalent to standard BN + SGD + Weight Decay + Momentum.
Exponential rate of convergence for harmonic heat flow maps.
problem Analyzing the convergence rate of harmonic heat flow maps.
method Proving exponential convergence rate for harmonic heat flow maps.
result Exponential convergence rate of the harmonic heat flow.
AdamNX improves Adam's stability by adjusting its learning rate.
problem Adam's tendency to converge to non-flat minima in large-scale models.
method Proposes a novel exponential decay mechanism for Adam's second-order moment estimate.
result AdamNX outperforms Adam and its variants in stability and performance.
Study shows exponential convergence in classification errors using random features and SGD.
problem Scalability issues in kernel methods for large datasets.
method Binary classification problem with random features and stochastic gradient descent.
result Exponential convergence rate of expected classification error achieved.
Extends likelihood ratio exponential families to analyze various optimization methods.
problem Analyzing optimization methods like rate-distortion and information bottleneck.
method Linking geometric mixture paths to exponential families and using hypothesis testing.
result Provides a common mathematical framework for understanding these methods.
Near-Exponential Convergence Rates for kNN Classification
problem Convergence rates for kNN classification
method Introducing Boltzmann margin
result First near-exponential convergence rates for kNN classification
MSGD outperforms SGD in overparametrized settings with faster convergence rates.
problem Optimization of non-convex functions with momentum.
method Momentum Stochastic Gradient Descent (MSGD) with rigorous analysis.
result MSGD converges exponentially faster than SGD in overparametrized settings.
SGD favors flat minima exponentially more than sharp minima in deep learning.
problem Understanding how SGD selects flat minima in deep learning.
method Developed a density diffusion theory (DDT) to analyze minima selection.
result SGD exponentially favors flat minima over sharp minima due to Hessian-dependent noise.
New ensemble method improves model stability exponentially.
problem Improving model stability for discontinuous base learners.
method Selecting the most frequently generated model from subsamples.
result Exponentially decaying tails for excess risk.
Deep neural networks approximate analytic functions in high dimensions with exponential rates.
problem Approximating analytic functions in high-dimensional spaces using neural networks.
method Analyzing convergence rates of ReLU and ReLU^k activations in L2(Rd,γd) for d∈N∪{∞}. result Exponential convergence rates for analytic functions in L2(Rd,γd) for d∈N, and dimension-independent bounds for d=∞. Boosting with tempered exponential measures improves AdaBoost's convergence rate.
problem Improving the convergence rate of AdaBoost.
method Introducing tempered exponential measures (TEMs) to generalize AdaBoost's approach.
result t-AdaBoost achieves an improved convergence rate compared to AdaBoost, especially for t∈[0,1). The AdaBoost algorithm was designed to combine many "weak" hypotheses that perform slightly better than random guessing into a "strong" hypothesis that has very low error. We study the rate at which AdaBoost iteratively converges to the minimum of the "exponential loss." Unlike previous work, our proofs do not require …
Develops European power option pricing under correlated interest rate and asset processes.
problem Pricing European power options under correlated interest rate and asset processes.
method Martingale method and Girsannov transform.
result Derives European power option pricing formulae under two market assumptions.
Paper proposes E/PD-Control for better neural network training.
problem Training efficiency and robustness of CNNs in online data flows.
method E/PD-Control combines feedback PD controller with exponential signal.
result Better learning efficiency and robustness demonstrated experimentally.
Study growth rates of subgroups in groups with a constricting element.
problem Understanding growth rates of subgroups in groups with a constricting element.
method Examining the spectrum of relative and quotient exponential growth rates of quasi-convex subgroups.
result Determine when growth rates of subgroups are strictly smaller or coincide with the group's growth rate.
Boosting improves data fitting while maintaining fairness guarantees.
problem Ensuring fairness in data preprocessing.
method Boosting algorithm to learn sufficient statistics of exponential families.
result The learned distribution maintains fairness guarantees while fitting the data better.
We provide a lower bound for the uniform exponential growth rate of closed nonflat nonpositively curved 3-manifold groups. A detailed study of the uniform exponential growth rate of closed 3-manifold groups is also presented.
Normalization layers control deep neural network capacity, improving stability and generalization.
problem Excessive capacity in deep neural networks leads to overfitting and poor generalization.
method Developed a theoretical framework to explain normalization's role in capacity control.
result Normalization layers reduce the Lipschitz constant exponentially, smoothing the loss landscape and enhancing generalization.
SGD with machine learning noise converges to global minimum exponentially fast.
problem Optimizing machine learning models with stochastic gradient descent.
method Analysis of SGD with machine learning noise, focusing on energy landscapes and gradient noise.
result SGD converges to the global minimum exponentially fast under certain conditions.
In our recent paper, we showed that in exponential family, contrastive divergence (CD) with fixed learning rate will give asymptotically consistent estimates \cite{wu2016convergence}. In this paper, we establish consistency and convergence rate of CD with annealed learning rate ηt. Specifically, suppose CD-m gener…
Study on convergence rate of Q-curvature flow in 6 dimensions.
problem Analyzing the convergence rate of Q-curvature flow in 6 dimensions. method Provided an example of a slowly converging Q6-curvature flow in dimension 6. result The Q-curvature flow in 6 dimensions does not always converge exponentially, unlike in 2 dimensions. Establishes exponential contraction in Wasserstein distance on manifolds and flows.
problem Analyzing contraction rates in Wasserstein distance on manifolds and their evolution.
method Explicit estimates and extension to evolving manifolds under geometric flow.
result Gradient estimates with exponential contraction rate under weak curvature conditions.
The study analyzes how machine learning classifiers' error rates decrease exponentially based on large deviations theory.
problem Understanding the convergence rate of machine learning classifiers' error probabilities.
method Large deviations theory applied to machine learning classification techniques.
result The error probability of ML classifiers converges to zero exponentially, with a rate dependent on the training set size.
Paper analyzes faster convergence rates for reinforcement learning from offline data.
problem Analyzing faster convergence rates for reinforcement learning from offline data.
method Fine analysis of reinforcement learning from offline data, providing fast rates for regret convergence.
result The paper provides fast rates for the regret convergence, showing that the level of exponentiation depends on the noise in the decision-making problem.
Paper establishes convergence rates and concentration bounds for stochastic approximation and reinforcement learning with Markovian noise.
problem Analyzing convergence rates and concentration bounds for stochastic approximation and reinforcement learning with Markovian noise.
method Novel discretization of the mean ODE of stochastic approximation algorithms using intervals with diminishing length.
result First almost sure convergence rate and maximal concentration bound with exponential tails for contractive stochastic approximation algorithms with Markovian noise.
The paper tackles fast rates in structured prediction problems.
problem Structured prediction problems with discrete outputs.
method Introducing continuous surrogate problems and leveraging their convergence rates for discrete problems.
result Super fast rates, including exponential rates, for excess risk in structured prediction problems.
MSTGD optimizes gradient descent with stratified sampling for faster convergence.
problem Fluctuation in gradient expectation and variance between iterations.
method Memory Stochastic Stratified Gradient Descent (MSTGD) with stratified sampling and variance reduction.
result MSTGD achieves an exponential convergence rate independent of dataset size and batch size.
Adapts to estimate functions from noisy ERT data.
problem Estimating functions from noisy Exponential Radon Transform data.
method Locally adaptive kernel type estimator for functions of varying smoothness.
result Achieves minimax optimal rate up to a log(n) factor for Sobolev functions.
Optimal learning for parametric prophet inequalities with exponential-type distributions
problem Learning in prophet inequalities with unknown parameters
method Confidence-based dynamic-programming policy
result Achieves optimal asymptotic competitive ratio using online observations
We prove that for analytic functions in low dimension, the convergence rate of the deep neural network approximation is exponential.
The paper explores how benign overfitting occurs in heavy-tailed input distributions.
problem Understanding overfitting in heavy-tailed input distributions.
method Analysis of maximum margin classifiers on unregularized logistic loss with gradient descent.
result Linear classifiers trained under certain conditions can asymptotically achieve the noise level as misclassification error.
This paper considers multi-dimensional affine processes with continuous sample paths. By analyzing the Riccati system, which is associated with affine processes via the transform formula, we fully characterize the regions of exponents in which exponential moments of a given process do not explode at any time or explode…
Kernel estimator optimally recovers function from noisy exponential Radon transform.
problem Inverting noisy exponential Radon transform of a function.
method Proposed a kernel estimator to estimate the true function.
result The estimator converges to the true function at minimax optimal rate.
Optimization rates improved for manifolds with bounded geometry.
problem Optimizing functions on manifolds with bounded geometry.
method Riemannian gradient descent and dynamic trivialization algorithm.
result Curvature-dependent convergence rates computed explicitly for common manifolds.
This paper studies a class of exponential family models whose canonical parameters are specified as linear functionals of an unknown infinite-dimensional slope function. The optimal minimax rates of convergence for slope function estimation are established. The estimators that achieve the optimal rates are constructed …
New algorithm trains deep neural networks without global optimization.
problem Training deep neural networks efficiently and without global optimization.
method Uses random complex exponential activation functions and Markov Chain Monte Carlo sampling.
result Consistently attains theoretical approximation rate for residual networks.
Study improves the exponential rate of metric difference in Higgs bundles.
problem Improving the exponential rate of metric difference in Higgs bundles.
method Analyzes the Hitchin metric and semi-flat metric in rank two Higgs bundles.
result Exponential rate of metric difference is improved.
Empirical study finds variance swap rate is affine in spot variance for S&P500 data.
problem Investigating the relationship between variance swap rate and spot variance.
method Empirical analysis using S&P500 data from 2006-2018, testing different models.
result Affine relationship between variance swap rate and spot variance is supported.
Active learning refers to the learning protocol where the learner is allowed to choose a subset of instances for labeling. Previous studies have shown that, compared with passive learning, active learning is able to reduce the label complexity exponentially if the data are linearly separable or satisfy the Tsybakov noi…
AdaX improves Adam by exponentially accumulating past gradients, leading to better performance in machine learning tasks.
problem Adam's fast convergence can lead to local minimums in non-convex problems.
method AdaX exponentially accumulates past gradients to adaptively tune the learning rate.
result AdaX outperforms Adam in various machine learning tasks, including computer vision and natural language processing.
Three-hidden-layer neural networks can approximate Hölder continuous functions uniformly with exponential rate.
problem Approximating Hölder continuous functions with neural networks.
method Introduced Floor-Exponential-Step (FLES) networks with three hidden layers.
result Uniform approximation of Hölder continuous functions with an exponential rate.