Research
On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

168,695 papers · 148 categories

Trend · papers per month

95190285380 · Jun 202019922001200920172026
48 results for decaying step sizes

Step decay schedules improve convergence in non-convex optimization.

problem Improving convergence in non-convex optimization problems.
method Analyzing convergence rates of step decay schedules in non-convex, convex, and strongly convex problems.
result Step decay schedules achieve O(lnT/T)\mathcal{O}(\ln T/\sqrt{T}) convergence rates in various optimization scenarios.

Paper develops an online learning algorithm for functional data models.

problem Recovering slope functions or predictors in functional data models.
method Online regularized learning algorithm in reproducing kernel Hilbert spaces with polynomially decaying step-size.
result Established fast convergence rates for estimation error without capacity assumption.

Improved analysis for fair federated learning reduces dependence on noise floor.

problem Asymptotic stationarity in group fair federated learning with reduced noise floor dependence.
method DS FedProxGrad framework with inexact local proximal solutions and fairness regularization.
result Algorithm converges asymptotically to stationarity without dependence on a noise floor.

In this paper, we study the online learning algorithm without explicit regularization terms. This algorithm is essentially a stochastic gradient descent scheme in a reproducing kernel Hilbert space (RKHS). The polynomially decaying step size in each iteration can play a role of regularization to ensure the generalizati…

2017-10-10abs ↗pdf ↗

We consider dd-dimensional linear stochastic approximation algorithms (LSAs) with a constant step-size and the so called Polyak-Ruppert (PR) averaging of iterates. LSAs are widely applied in machine learning and reinforcement learning (RL), where the aim is to compute an appropriate θRdθ_{*} \in \mathbb{R}^d (that is a…

2017-09-12abs ↗pdf ↗

Proposes an exponentially increasing step-size for faster parameter estimation in statistical models.

problem Slow convergence of gradient descent in locally convex loss functions.
method Exponentially increasing step-size in gradient descent algorithm.
result Converges linearly to optimal solution under homogeneous assumptions.

Riemannian stochastic gradient descent converges faster with increasing batch size.

problem Improving convergence rate of Riemannian stochastic gradient descent.
method Theoretical analysis and numerical investigation of increasing batch size effects.
result Riemannian stochastic gradient descent converges faster with increasing batch size.

AdamP optimizes momentum-based optimizers for scale-invariant weights, improving model performance.

problem Premature decay of effective step sizes in momentum-based optimizers for scale-invariant weights.
method Proposes SGDP and AdamP to eliminate the radial component at each optimizer step, preserving convergence properties.
result Uniform gains across multiple benchmarks, improving model performance.

Seesaw optimizes training by balancing learning rate and batch size, accelerating model pretraining.

problem Optimizing training efficiency for large language models with adaptive optimizers.
method Develops a principled framework for batch-size scheduling, introducing Seesaw which multiplies learning rate by 1/√2 and doubles batch size.
result Empirically, Seesaw reduces wall-clock time by approximately 36% compared to cosine decay, matching theoretical limits.

SignSGD outperforms SGD in linear regression with optimal scaling laws under PLRF model.

problem Improving linear regression performance with signSGD under power-law random features.
method Analysis of signSGD risk under PLRF model, comparison with SGD, identification of unique effects.
result SignSGD can have a steeper compute-optimal slope than SGD in noisy regimes, especially with WSD schedule.

Gradient descent outperforms ridge regression under certain covariance matrix decay conditions.

problem Comparing the performance of gradient descent and ridge regression in linear models.
method Investigated gradient descent and ridge regression for linear regression with random isotropic ground truth.
result Gradient descent outperforms ridge regression under specific covariance matrix decay conditions.

AMAGOLD improves stochastic gradient MCMC by infrequent Metropolis-Hastings corrections.

problem Bias in stochastic gradient Hamiltonian Monte Carlo (SGHMC).
method AMAGOLD infrequently uses Metropolis-Hastings corrections to remove bias, with a fixed step size schedule.
result AMAGOLD converges to the target distribution with a fixed, rather than a diminishing, step size, and at most a constant factor slower convergence rate.

Paper analyzes SGD in kernel regression, showing it outperforms offline methods.

problem Performance of SGD in kernel regression compared to offline methods.
method Analyzes Stochastic Gradient Descent (SGD) in kernel regression under misspecified models.
result SGD achieves min-max optimal rates up to constants, avoiding saturation.

New theory sharpens Q-learning with LDTZ rate, proving it's best of both worlds.

problem Improving Q-learning's theoretical and practical performance.
method Developed a sharp non-asymptotic error bound and central limit theory for Q-learning with PD2Z-ν schedule.
result Q-learning with LDTZ schedule achieves rapid decay and asymptotic convergence guarantees.

SGDm with fixed step-size diverges under covariate shift, similar to a parametric oscillator.

problem SGDm with fixed step-size diverges under covariate shift.
method Approximated learning system as a time-varying system of ODEs and characterized divergence/convergence modes.
result SGDm with fixed step-size can diverge under covariate shift, similar to resonance in oscillators.

Study convergence of simulated annealing in continuous and discrete settings.

problem Analyzing convergence rate of simulated annealing methods.
method Apply Eyring-Kramers law to prove polynomial decay of tail probabilities.
result Explicit rate of convergence for continuous and discrete simulated annealing.

A random walk wnw_n on a separable, geodesic hyperbolic metric space XX converges to the boundary X\partial X with probability one when the step distribution supports two independent loxodromics. In particular, the random walk makes positive linear progress. Progress is known to be linear with exponential decay when …

2017-10-14abs ↗pdf ↗

Adam's generalization performance is improved by batch size and weight decay in neural networks.

problem Understanding how batch size and weight decay affect Adam's generalization in neural networks.
method Theoretical analysis of two-layer over-parameterized CNNs on image data.
result Adam's mini-batch variants can achieve near-zero test error, unlike full-batch Adam.

Stochastic variance reduction algorithms have recently become popular for minimizing the average of a large, but finite, number of loss functions. In this paper, we propose a novel Riemannian extension of the Euclidean stochastic variance reduced gradient algorithm (R-SVRG) to a compact manifold search space. To this e…

2016-05-24abs ↗pdf ↗

In this paper, we consider unregularized online learning algorithms in a Reproducing Kernel Hilbert Spaces (RKHS). Firstly, we derive explicit convergence rates of the unregularized online learning algorithms for classification associated with a general gamma-activating loss (see Definition 1 in the paper). Our results…

2015-03-02abs ↗pdf ↗

Novel bounds for SGLD show generalization error decreases with more samples.

problem Understanding the generalization error of SGLD in non-convex optimization.
method Information-theoretic approach focusing on Kullback-Leibler divergence and sub-exponential loss function.
result Time-independent generalization bounds for SGLD, independent of step size and number of iterations.

Stagewise training strategy is widely used for learning neural networks, which runs a stochastic algorithm (e.g., SGD) starting with a relatively large step size (aka learning rate) and geometrically decreasing the step size after a number of iterations. It has been observed that the stagewise SGD has much faster conve…

2018-12-10abs ↗pdf ↗

A new algorithm for decentralized optimization over directed graphs.

problem Decentralized stochastic optimization over directed networks.
method Gradient tracking and S-ADDOPT algorithm with constant and decaying step-sizes.
result S-ADDOPT converges linearly with constant step-size and sublinearly with decaying step-size.

Study shows splitting schemes can approximate WFR flows faster than the exact flow.

problem Improving sampling efficiency in Wasserstein-Fisher-Rao gradient flows.
method Investigates operator splitting techniques to numerically approximate WFR flows.
result A judicious choice of step size and operator ordering can lead to faster convergence of split schemes to the target distribution.

Improved convergence for actor-critic algorithms in MDPs.

problem Global convergence analysis for actor-critic algorithms in MDPs.
method Introduced an analytical framework to handle complex recursions, established convergence to ε-close globally optimal policy with improved sample complexity.
result Converges to ε-close globally optimal policy with sample complexity of O(ε^(-3)) compared to O(ε^(-2)) for ε-close stationary policy.

Approximations to utility indifference prices are provided for a contingent claim in the large position size limit. Results are valid for general utility functions on the real line and semi-martingale models. It is shown that as the position size approaches infinity, the utility function's decay rate for large negative…

2012-02-17abs ↗pdf ↗

New findings show convergence in SA without square summable step sizes.

problem Finding optimal step sizes for stochastic approximation algorithms.
method Analyzed step-size sequences of the form αn=α0nρα_n = α_0 n^{-ρ} with ρ(0,1)ρ\in (0,1), and derived convergence and rate results.
result Convergence of parameter estimates with probability one and in LpL_p for any p1p\ge 1; MSE rate of convergence is O(αn)O(α_n), improved to O(max{αn2,1/n})O(\max\{ α_n^2,1/n \}) with averaging.

Study on RF regression with SGD shows double descent phenomenon.

problem Understanding generalization in RF models trained with SGD.
method Precise non-asymptotic error bounds derived for RF regression under constant and polynomial-decay step-size SGD.
result RF regression generalizes well for interpolation learning and exhibits double descent behavior.

The paper analyzes reg-SGD for convex problems, proving convergence and quantifying the rate of convergence.

problem Minimizing convex, L-smooth functions in a Hilbert space.
method Regularized stochastic gradient descent with decaying regularization.
result Strong convergence to the minimum-norm solution without boundedness assumptions.

Temporal Difference Learning analysis under non-i.i.d. data and nonlinear approximation.

problem Finite-sample behavior of TD(0) under non-i.i.d. data and nonlinear approximation.
method High-probability, finite-sample analysis of vanilla TD(0) on polynomially mixing Markov data, assuming Holder continuity and bounded generalized gradients.
result Bounds on the convergence rate of TD(0) with high probability, matching known i.i.d. rates and holding even with nonstationary initialization.

Implicit Q-learning and SARSA adjust step-sizes automatically, improving stability and performance.

problem Numerical instability and slow progress in Q-learning and SARSA due to step-size calibration.
method Reformulate iterative updates as fixed-point equations, scaling step-sizes inversely with feature norms.
result Implicit methods maintain stability over broader step-size ranges and achieve comparable convergence rates.

The study introduces anytime learning schedules for large language models without fixed horizons.

problem Training large language models without knowing the total training horizon.
method Theoretical analysis and weight averaging to create anytime learning schedules.
result Theoretical and empirical evidence shows that weight averaging with simple step sizes can achieve comparable final loss to well-tuned cosine schedules.

Stochastic momentum methods trade compute efficiency for serial runtime.

problem Stochastic momentum methods trade compute efficiency for serial runtime.
method Stochastic HB and ASGD for consistent linear regression with Gaussian covariates.
result HB preserves SGD-level CE over a larger batch-size window, allowing larger batches to reduce serial runtime until HB reaches its deterministic accelerated scale.

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.