Dual SVM training with budget constraint for faster accuracy.
problem Efficient support vector machine training with limited resources.
method Dual subspace ascent algorithm with budget constraint.
result Significant speed-up over primal budget training methods.
Studying a softmax-attention model, we show that the learned query converges to the latent signal subspace spanned by the informative direction.
problem Understanding the theoretical principles of attention mechanisms in large-scale token collections.
method Deriving a population objective and analyzing the limiting ordinary differential equation of the learning dynamics.
result The learned query asymptotically recovers the latent signal up to the intrinsic sign ambiguity.
SMAVE optimizes SDR by projecting onto a low-dimensional subspace on a Riemannian manifold.
problem High-dimensional regression challenges due to the curse of dimensionality.
method SMAVE combines nearest-neighbor localization and Riemannian stochastic gradient ascent.
result SMAVE achieves almost-sure convergence and matches RMAVE's synthetic subspace recovery rate.
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.
Develops an MS-inspired algorithm for regression mode finding and space partitioning.
problem Finding local modes of regression functions and partitioning input space.
method Mean-shift-inspired algorithm for iterative gradient ascent.
result Proves convergence and rates of convergence for estimated local modes.
This paper presents a stochastic behavior analysis of a kernel-based stochastic restricted-gradient descent method. The restricted gradient gives a steepest ascent direction within the so-called dictionary subspace. The analysis provides the transient and steady state performance in the mean squared error criterion. It…
New algorithm for active exploration in bandit problems.
problem Active exploration in bandit problems.
method Gradient ascent with online lazy mirror ascent sampling rule.
result Asymptotically optimal and computationally efficient.
Improves posterior approximation speed for Dirichlet process mixture models.
problem Inefficiency of stochastic variational inference in large datasets.
method Uses stochastic gradient ascent with adaptive stepsize optimization.
result Adaptive stepsize improves speed and performance of posterior approximation.
Gradient ascent method successfully removes specific data points from neural networks without retraining.
problem Addressing privacy and ethical concerns by removing specific data points from trained models.
method Gradient ascent approach to unlearning, leveraging the implicit bias of gradient descent towards margin maximization conditions.
result Gradient ascent method can successfully unlearn specific data points from two-layer ReLU neural networks without retraining.
Sequential coordinate ascent is more robust in high-dimensional linear regression.
problem Behavior difference between sequential and parallel coordinate ascent in variational inference.
method Comparison of sequential and parallel coordinate ascent algorithms in high-dimensional linear regression.
result Sequential algorithm converges under more relaxed conditions than parallel algorithm.
Optimal Coordinate Ascent (OCA) improves feature selection in machine learning.
problem Efficiently selecting features in machine learning models.
method Optimal Coordinate Ascent (OCA) for feature selection.
result OCA outperforms previous methods in feature selection and model performance.
Stochastic Gradient Descent (SGD) has become popular for solving large scale supervised machine learning optimization problems such as SVM, due to their strong theoretical guarantees. While the closely related Dual Coordinate Ascent (DCA) method has been implemented in various software packages, it has so far lacked go…
Stochastic dual coordinate ascent (SDCA) is an effective technique for solving regularized loss minimization problems in machine learning. This paper considers an extension of SDCA under the mini-batch setting that is often used in practice. Our main contribution is to introduce an accelerated mini-batch version of SDC…
New algorithm solves federated minimax optimization problems.
problem Federated minimax optimization challenges.
method Federated Stochastic Smoothed Gradient Descent Ascent (FESS-GDA).
result FESS-GDA uniformly solves federated minimax problems.
We introduce a proximal version of dual coordinate ascent method. We demonstrate how the derived algorithmic framework can be used for numerous regularized loss minimization problems, including ℓ 1 \ell_1 ℓ 1 regularization and structured output SVM. The convergence rates we obtain match, and sometimes improve, state-of-the-…
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.
Policy gradient aims to maximize expected return using gradient ascent.
problem Finding a policy that maximizes expected return in a given class of policies.
method Gradient ascent applied to a differentiable model of the policy, estimating the gradient of expected return.
result Policy gradient methods require on-policy data for gradient estimation, limiting sample efficiency.
Algorithm improves variational inference in Wasserstein distance.
problem Improving variational inference methods for complex models.
method Wasserstein contraction analysis of coordinate ascent.
result General and sharp convergence guarantees for various models.
In this paper we study several classes of stochastic optimization algorithms enriched with heavy ball momentum. Among the methods studied are: stochastic gradient descent, stochastic Newton, stochastic proximal point and stochastic dual subspace ascent. This is the first time momentum variants of several of these metho…
Paper introduces SGA for barycenter optimization in optimal transport.
problem Optimizing Wasserstein barycenter for discrete distributions.
method Sobolev gradient ascent algorithm tailored to Wasserstein geometry.
result SGA achieves convergence rate similar to subgradient descent.
Gradient descent-ascent converges to strict local minmax equilibria with a finite timescale separation.
problem Analyzing the convergence of gradient descent-ascent in non-convex, non-concave games with a finite timescale separation.
method Investigates the role of a finite timescale separation parameter τ on gradient descent-ascent in two-player zero-sum games, providing convergence rates and non-convergence results.
result Gradient descent-ascent converges to strict local minmax equilibria for a finite timescale separation parameter τ*.
LORENZA improves LLM fine-tuning efficiency and generalization.
problem Improving robustness and generalization of LLMs under hardware constraints.
method AdaZo-SAM and LORENZA, combining Adam and SAM with zeroth-order estimation and randomized SVD.
result LORENZA achieves better generalization and reduced memory consumption compared to existing methods.
Lazy neural networks are vulnerable to adversarial attacks.
problem Adversarial robustness issues in lazy training models.
method Extending recent work on adversarial examples to lazy training models.
result Over-parametrized neural networks that generalize well remain vulnerable to single-step gradient ascent attacks.
Policy mirror ascent achieves Nash equilibrium in mean field games without a population generative model.
problem Achieving Nash equilibrium in mean field games without a population generative model.
method Policy mirror ascent, contractive operator, single-path TD learning.
result Policy mirror ascent converges to Nash equilibrium within O ~ ( ε − 2 ) \widetilde{\mathcal{O}}(\varepsilon^{-2}) O ( ε − 2 ) samples. New insights into gradient descent and ascent dynamics in min-max optimization.
problem Understanding the convergence and limit points of gradient descent and ascent methods in min-max optimization problems.
method Characterization of limit points using dynamical systems perspective for GDA and OGDA.
result Both GDA and OGDA dynamics avoid unstable critical points and have a superset of local min-max solutions.
Paper proposes an algorithm to solve complex minimax problems efficiently.
problem Stochastic nonconvex-concave minimax problems in various fields.
method Accelerated first-order regularized momentum descent ascent algorithm (FORMDA).
result Achieves best-known complexity bound of i l d e O ( ε − 6.5 ) ilde{\mathcal{O}}(\varepsilon ^{-6.5}) i l d e O ( ε − 6.5 ) for single-loop algorithms. The stochastic dual coordinate-ascent (S-DCA) technique is a useful alternative to the traditional stochastic gradient-descent algorithm for solving large-scale optimization problems due to its scalability to large data sets and strong theoretical guarantees. However, the available S-DCA formulation is limited to finit…
Random scan CAVI converges linearly under log-concave assumptions.
problem Analyzing the convergence rate of random scan Coordinate Ascent Variational Inference (CAVI) under log-concave conditions.
method Building on previous work, we analyze the random scan version of CAVI using optimal transport geometry.
result We obtain tight linear convergence rates for the random scan version of CAVI.
Unified analysis of EG and OGDA for saddle point problems using proximal point method.
problem Solving saddle point problems in bilinear and strongly convex-strongly concave settings.
method Unified analysis as approximations of the proximal point method.
result Unified analysis of EG and OGDA for saddle point problems.
We introduce a proximal version of the stochastic dual coordinate ascent method and show how to accelerate the method using an inner-outer iteration procedure. We analyze the runtime of the framework and obtain rates that improve state-of-the-art results for various key machine learning optimization problems including …
Paper tackles online PCA, showing non-convex gradient ascent can minimize regret.
problem Online Principal Component Analysis in adversarial settings.
method Adversarially-perturbed spiked-covariance model, non-convex online gradient ascent with warm-start.
result Non-convex online gradient ascent can minimize regret with high probability.
The paper analyzes inexact variants of iterative methods for solving optimization problems.
problem Solving optimization problems with inexact sub-problems.
method Inexact variants of stochastic gradient descent, Newton, proximal point, and subspace ascent methods.
result Iteration complexity results for inexact variants of various iterative methods.
Gradient-descent-ascent dynamics can exhibit various behaviors in non-convex non-concave games.
problem Gradient-descent-ascent dynamics in non-convex non-concave games can lead to recurrent behavior and spurious equilibria.
method Combines optimization theory, game theory, and dynamical systems.
result Gradient-descent-ascent dynamics can exhibit Poincaré recurrence and converge to spurious equilibria.
New taxonomy and improved solvers for discrete energy minimization.
problem Maximum-a-posteriori inference in discrete graphical models.
method Dual block-coordinate ascent rule, theoretical analysis, new solver variants.
result Improved state-of-the-art solver outperforming existing methods on all test instances.
Optimizes binary regression models with gradient ascent-descent methods.
problem Regression problems with binary weights in quantized learning and digital communication.
method Maximin optimization using gradient ascent-descent methods.
result The approach is optimal in linear regression with low noise and robust regression with few outliers.
New ODE models show saddle-point optimization methods converge differently, with last-iterate convergence for OGDA.
problem Analyzing convergence properties of saddle-point optimization methods.
method High-Resolution Differential Equations (HRDEs) to design differential equation models for saddle-point optimization methods.
result HRDEs reveal last-iterate convergence for Optimistic Gradient Descent Ascent (OGDA) in bilinear games.
Develops a new framework for analyzing MFVI algorithms.
problem Analyzes mean field variational inference (MFVI) formulations.
method Inspired by variational Bayesian formulations, represents MFVI problem in three ways: gradient flow, Fokker-Planck-like equations, and diffusion process.
result Establishes rigorous guarantees for convergence of time-discretized coordinate ascent variational inference algorithms.
Paper studies convergence of Mean-Field GDA dynamics for MNE of continuous games.
problem Finding mixed Nash equilibria in continuous games.
method Two-scale Mean-Field Gradient Descent Ascent dynamics.
result Two-scale Mean-Field GDA converges exponentially to MNE without convexity assumptions.
The study examines statistical inference with gradient ascent in multi-modal likelihood functions.
problem Statistical inference with multiple initializations in multi-modal likelihood functions.
method Derives population quantity, studies asymptotic normality, bootstrap, and likelihood ratio tests.
result Coverage deficiency and differences in CIs due to finite number of initializations.
This paper introduces a new method to evaluate multiple policies simultaneously.
problem Estimating the value of many policies for a single set of states.
method Developed a scalable, differentiable fingerprinting mechanism to represent complex policies.
result The method can produce policies that outperform those that generated the training data, in zero-shot manner.
SAVO actor improves reinforcement learning by avoiding local optima in complex Q-functions.
problem Gradient ascent in complex Q-functions leads to suboptimal solutions.
method SAVO actor generates multiple action proposals and truncates poor local optima.
result SAVO actor finds optimal actions more frequently and outperforms other architectures.
We introduce the notion of ascent sliceness of virtual knots. A representative of a virtual knot is an embedding S 1 ↪ Σ g × I S^1 \hookrightarrow Σ_{g} \times I S 1 ↪ Σ g × I , for Σ g Σ_g Σ g a closed connected oriented surface of genus g g g ; the virtual knot represented is slice if there exists a pair consisting of a disc D D D and an oriented…
This paper introduces AdaSDCA: an adaptive variant of stochastic dual coordinate ascent (SDCA) for solving the regularized empirical risk minimization problems. Our modification consists in allowing the method adaptively change the probability distribution over the dual variables throughout the iterative process. AdaSD…
Algorithm converges to Nash equilibria in competitive games.
problem Finding Nash equilibria in decentralized, competitive Markov games.
method Decentralized Optimistic Gradient Descent/Ascent with a critic.
result Converges to the set of Nash equilibria under self-play.
The paper connects two clustering methods by showing gradient ascent flow can move up the cluster tree.
problem Establishing a strong correspondence between clustering methods.
method Moving up the cluster tree by following the gradient ascent flow.
result Gradient ascent flow can be used to move up the cluster tree.
Gradient Descent Ascent converges to von-Neumann solution in hidden zero-sum games.
problem Understanding dynamics of zero-sum games with hidden structure.
method Gradient Descent Ascent applied to hidden zero-sum games with specific convex-concave structure.
result Gradient Descent Ascent converges to von-Neumann solution in strictly convex-concave hidden games.
Paper proposes algorithms for solving nonconvex-nonconcave problems with complexity guarantees.
problem Nonconvex-nonconcave minimax problems with PL condition.
method Zeroth-order AGDA and VRAGDA algorithms.
result Iteration complexities for obtaining ε-stationary points.
SGD finds global optima in WGANs for 1-layer generators.
problem Training GANs with WGANs requires global optimality, which is hard.
method Used SGD to train 1-layer generator networks in WGANs.
result SGD converges to global solution in polynomial time and sample complexity.