Paper proves linear convergence of SCMS algorithm for directional data.
problem Identifying density ridges in directional data.
method Generalized SCMS algorithm to directional data, derived from SCGA with adaptive step size.
result Linear convergence of the proposed directional SCMS algorithm.
Paper finds conditions for benign overfitting in neural networks.
problem Benign overfitting in leaky ReLU two-layer neural networks.
method Established directional convergence and classification error bounds.
result Benign overfitting occurs with high probability on mixture data.
Push-SAGA is a decentralized algorithm for directed graphs that converges linearly.
problem Finite-sum minimization over directed graphs with stochastic gradients.
method Combines variance reduction, gradient tracking, and consensus algorithms.
result Achieves linear convergence for smooth and strongly convex problems.
Study shows directional convergence for neural networks under spherical symmetry.
problem Learning linear predictors with neural networks under spherically symmetric data.
method Analysis of gradient flow and gradient descent for two-layer and deep linear networks.
result Directional convergence guarantees with exact convergence rate for specific network architectures.
New stochastic gradient descent with random search directions improves efficiency and convergence.
problem Efficiency and convergence of stochastic gradient descent methods.
method Developed a new class of stochastic gradient descent algorithms with random search directions.
result Established almost sure convergence and provided Lp rates of convergence. Yau's Affine Normal Descent optimizes smooth unconstrained problems with geometrically adapted directions.
problem Optimizing smooth unconstrained problems with geometrically adapted directions.
method Yau's Affine Normal Descent (YAND) uses the equi-affine normal of level-set hypersurfaces as search directions.
result YAND converges globally under standard smoothness assumptions and locally quadratically near nondegenerate minimizers.
The paper studies kernel smoothing and mean shift for directional data, deriving convergence rates and mode estimation.
problem Statistical and computational problems of kernel smoothing for directional data.
method Generalization of mean shift to directional data, derivation of convergence rates, and investigation of mode estimation.
result Statistical convergence rates of directional KDE and its derivatives, ascending property of directional mean shift, and mode estimation.
Paper shows DMS as an EM algorithm with improved convergence.
problem Improving the convergence of DMS algorithm.
method Shows DMS as a generalized EM algorithm and provides new proofs.
result Demonstrates global convergence and linear convergence of DMS.
Direct method finds Yang-Mills connections for SO(3) bundles.
problem Finding Yang-Mills connections for SO(3) bundles over closed 4-manifolds.
method Direct minimizing method with test connections and assumptions.
result Minimizing sequence converges to a minimizer or anti-selfdual/selfdual connection.
The paper studies neural networks' convergence near origin and saddle points.
problem Directional convergence of neural networks near small initializations and saddle points.
method Gradient flow dynamics analysis of two-homogeneous neural networks.
result Neural networks' weights approximately converge in direction to KKT points for small initializations.
Deep networks converge in direction, with implications for predictions and margins.
problem Understanding convergence and alignment in deep learning networks.
method Developed a theory of unbounded nonsmooth Kurdyka-Łojasiewicz inequalities for functions definable in an o-minimal structure.
result Network weights, predictions, training errors, and margin distribution converge in direction and align with gradient flow.
Gradient descent implicitly follows regularization for general losses.
problem The implicit bias of gradient descent methods in machine learning.
method Empirical risk minimization over linear predictors with arbitrary convex, strictly decreasing losses.
result Gradient descent and regularization paths converge to the same direction for non-attained risks.
AB-SAGA optimizes distributed optimization over directed graphs using variance reduction and stochastic weights.
problem Optimizing distributed stochastic optimization over directed graphs with stochastic weights.
method AB-SAGA combines variance reduction and network-level gradient tracking, using both row and column stochastic weights.
result AB-SAGA converges linearly to the global optimal with a constant step-size and achieves a linear speed-up over centralized methods.
ConMeZO speeds up zeroth-order optimization for large language models.
problem Slow convergence in high-dimensional parameter spaces of large language models.
method Adaptive directional sampling in a cone centered around a momentum estimate.
result Achieves the same convergence rate as MeZO but up to 2X faster.
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.
Early training of deep neural networks leads to small, directionally converging weights.
problem Training dynamics of deep homogeneous neural networks with small initializations.
method Gradient flow analysis and study of KKT points for neural correlation function.
result Weights converge in direction to KKT points during early training stages.
Paper confirms Thom's conjecture for nonlinear evolutions on manifolds.
problem Thom's gradient conjecture for nonlinear evolution equations.
method Extending and settling the conjecture in infinite dimensional problems using Łojasiewicz, L. Simon, and Kurdyka-Mostowski-Parusinski's foundational works.
result Uniqueness of the limiting direction and characterization of convergence rates for both classical and infinite dimensional settings.
Enhances deep learning by boosting generalization and convergence.
problem Improving generalization and convergence in deep learning models.
method Implicit Regularization Enhancement (IRE) framework that decouples flat and sharp directions.
result IRE consistently improves generalization performance across various deep learning tasks and models.
We introduce a deformation of Riemann surfaces and we are interested in the convergence of this deformation to a point of the Gardiner-masur boundary of Teichmueller space. This deformation, which we call the horocyclic deformation, is directed by a projective measured foliation and belongs to a certain horocycle in a …
In optimization, the negative gradient of a function denotes the direction of steepest descent. Furthermore, traveling in any direction orthogonal to the gradient maintains the value of the function. In this work, we show that these orthogonal directions that are ignored by gradient descent can be critical in equilibri…
We study the implicit bias of gradient descent methods in solving a binary classification problem over a linearly separable dataset. The classifier is described by a nonlinear ReLU model and the objective function adopts the exponential loss function. We first characterize the landscape of the loss function and show th…
SGD benefits from a directional bias in kernel regression models.
problem Improving generalization in kernel regression models.
method Generalized directional bias property of SGD in kernel regression.
result SGD converges along the eigenvector of the largest eigenvalue of the Gram matrix.
Direct proof shows adaptive gradient descent converges near-linearly for convex functions.
problem Proving near-linear convergence of adaptive gradient descent for convex functions.
method Direct Lyapunov-based argument for convex functions with unique minimizer.
result Direct proof of near-linear convergence for convex functions.
Reconstructing polytopes with fixed facet directions from support function evaluations.
problem Reconstructing polytopes with known facet directions from limited data.
method Least-squares estimate via convex quadratic program, combinatorial characterization for uniqueness, algorithm convergence.
result The least-squares estimate for a fixed simplicial normal fan is a convex quadratic program, and the solution is unique under certain conditions.
In this paper, we consider solving a class of nonconvex and nonsmooth problems frequently appearing in signal processing and machine learning research. The traditional alternating direction method of multipliers encounters troubles in both mathematics and computations in solving the nonconvex and nonsmooth subproblem. …
We present a distributed (non-Bayesian) learning algorithm for the problem of parameter estimation with Gaussian noise. The algorithm is expressed as explicit updates on the parameters of the Gaussian beliefs (i.e. means and precision). We show a convergence rate of O(1/k) with the constant term depending on the numb…
Interneurons improve learning in neural networks by accelerating convergence.
problem Rapid adaptation to changing input statistics in neural networks.
method Two mathematically tractable recurrent linear neural networks were compared: one with direct recurrent connections and the other with interneurons that mediate recurrent communication.
result The network with interneurons converges more quickly than the network with direct recurrent connections, scaling logarithmically with initialization spectrum.
GD at EoS edge minimizes logistic loss without monotonic convergence.
problem Understanding GD's implicit bias at the edge of stability.
method Theoretical analysis of logistic regression with constant stepsize GD.
result GD with any constant stepsize minimizes logistic loss over long time scales.
We revisit the classical Douglas-Rachford (DR) method for finding a zero of the sum of two maximal monotone operators. Since the practical performance of the DR method crucially depends on the stepsizes, we aim at developing an adaptive stepsize rule. To that end, we take a closer look at a linear case of the problem a…
In this paper, we design and analyze a new zeroth-order online algorithm, namely, the zeroth-order online alternating direction method of multipliers (ZOO-ADMM), which enjoys dual advantages of being gradient-free operation and employing the ADMM to accommodate complex structured regularizers. Compared to the first-ord…
A new hybrid-ordered SGD method reduces communication and complexity for non-convex optimization.
problem Balancing communication, computational complexity, and convergence rate in distributed non-convex optimization.
method Hybrid-ordered distributed SGD with pre-shared scalers and periodic vector communication.
result Order-wise faster convergence compared to existing methods.
Improved SSD for faster and more accurate goodness-of-fit tests and model learning.
problem Optimal slicing directions for SSD are computationally expensive and sub-optimal.
method Relaxed optimal slicing requirement, active sub-space construction, spectral decomposition.
result 14-80x speed-up in goodness-of-fit tests compared to gradient-based alternatives.
Gradient descent, when applied to the task of logistic regression, outputs iterates which are biased to follow a unique ray defined by the data. The direction of this ray is the maximum margin predictor of a maximal linearly separable subset of the data; the gradient descent iterates converge to this ray in direction a…
Stochastic Gradient Descent shows directional bias with moderate learning rates, impacting optimization outcomes.
problem Understanding the bias of SGD with moderate learning rates in practical scenarios.
method Analyzing SGD and GD on an overparameterized linear regression problem.
result SGD converges along large eigenvalue directions, GD along small ones, affecting early stopping outcomes.
We propose a new stochastic dual coordinate ascent technique that can be applied to a wide range of regularized learning problems. Our method is based on Alternating Direction Multiplier Method (ADMM) to deal with complex regularization functions such as structured regularizations. Although the original ADMM is a batch…
We propose two new alternating direction methods to solve "fully" nonsmooth constrained convex problems. Our algorithms have the best known worst-case iteration-complexity guarantee under mild assumptions for both the objective residual and feasibility gap. Through theoretical analysis, we show how to update all the al…
Weyl's intrinsic volumes converge to the Euler characteristic of the base manifold under certain metrics.
problem Convergence of intrinsic volumes on Riemannian manifolds.
method Defined a new metric and used it to study the convergence of intrinsic volumes.
result Intrinsic volumes converge to the Euler characteristic of the base manifold.
We find the explicit expression for the equilibrium wealth distribution of the Directed Random Market process, recently introduced by Martínez-Martínez and López-Ruiz, which turns out to be a Gamma distribution with shape parameter 21. We also prove the convergence of the discrete-time process describing the…
Paper proposes PPMM for fast estimation of large-scale OTM.
problem Estimation of large-scale optimal transport maps (OTM) is challenging due to the curse of dimensionality.
method Combines projection pursuit regression and sufficient dimension reduction to adaptively select projection directions.
result PPMM consistently estimates the most informative projection direction and weakly converges to the target OTM.
Proposes new stochastic algorithms for multi-objective optimization.
problem Multi-objective optimization in machine learning problems.
method Direction-oriented multi-objective formulation and Stochastic Direction-oriented Multi-objective Gradient descent (SDMGrad).
result Stochastic algorithms converge to Pareto stationary points with improved complexities.
Study on slow convergence in geometric variational problems.
problem Slow convergence of solutions in geometric variational problems.
method Identifying necessary conditions for slowly converging solutions and characterizing their convergence rate and direction.
result Characterization of the rate and direction of convergence for slowly converging solutions.
This paper analyzes quantiles of heavy-tailed distributions, separating projection direction and quantile threshold effects.
problem Analyzing quantiles of heavy-tailed distributions with estimated parameters.
method Introduces a Q-Q orthogonality formulation to separate projection-direction and quantile-threshold effects.
result Decomposes the difference between empirical and population quantiles into three terms.
New algorithm optimizes nonlinear SDEs online with convergence guarantees.
problem Optimizing nonlinear stochastic differential equations (SDEs) is computationally challenging.
method Forward propagation algorithm that solves an SDE derived using forward differentiation.
result Convergence theorem for nonlinear dissipative SDEs with bounds on stochastic fluctuations.
We study the convergence of earthquake paths and horocycle paths in the Gardiner-Masur compactification of Teichmüller space. We show that an earthquake path directed by a uniquely ergodic or simple closed measured geodesic lamination converges to the Gardiner-Masur boundary. Using the embedding of flat metrics into th…
Proves convergence of gradient Ricci shrinkers with uniform bounds.
problem Compactness and energy concentration in gradient Ricci shrinkers.
method Bubble-tree convergence and local energy analysis.
result No energy concentrates in neck regions, leading to a local diffeomorphism finiteness theorem.
This paper analyzes convergence of FL for neural networks using NTK.
problem Theoretical guarantees of FL for neural networks with explicit forms and multi-step updates are unexplored.
method FL-NTK framework for federated learning of ReLU neural networks trained by gradient descent.
result FL-NTK converges to a global-optimal solution at a linear rate with proper learning parameters.
New algorithms estimate Hessians using random directions for faster stochastic optimization.
problem Efficiently estimating Hessians for stochastic optimization.
method Generalized Hessian estimators using random directions and noisy function measurements.
result Asymptotically unbiased estimators with lower bias for more measurements.
A recent algorithmic family for distributed optimization, DIGing's, have been shown to have geometric convergence over time-varying undirected/directed graphs. Nevertheless, an identical step-size for all agents is needed. In this paper, we study the convergence rates of the Adapt-Then-Combine (ATC) variation of the DI…