OMWU shows last iterate convergence in convex-concave games.
problem Optimizing in constrained min-max optimization landscapes.
method OMWU (Optimistic Multiplicative-Weights Update) in the no-regret online learning framework.
result OMWU exhibits last iterate convergence for convex-concave games, generalizing previous results.
New algorithms prove fast convergence in complex min-max problems.
problem Proving fast convergence in nonconvex min-max optimization.
method Hamiltonian Gradient Descent (HGD) and Consensus Optimization (CO) algorithms.
result HGD and CO achieve linear convergence in various settings.
Improved convergence rates for saddle-point optimization algorithms.
problem Understanding last-iterate convergence rates for saddle-point optimization algorithms in constrained settings.
method Expanding the understanding of last-iterate convergence for Optimistic Gradient Descent Ascent (OGDA) and Optimistic Multiplicative Weights Update (OMWU) in the constrained setting.
result Linear last-iterate convergence achieved with a universal constant learning rate for OMWU in bilinear games over the simplex.
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.
Policy gradient method proves convergence in imperfect-information games.
problem Policy gradient methods in imperfect-information games (EFGs).
method Policy gradient approach with best-iterate convergence.
result Policy gradient leads to provable best-iterate convergence in self-play EFGs.
This work addresses the convergence of SGD's final iterate without restrictive assumptions.
problem Prove optimal convergence rate of SGD's final iterate without compact domains or bounded noise.
method Unified proof for general domains, composite objectives, non-Euclidean norms, etc.
result First unified convergence rates in expectation and high probability.
Efficient algorithm converges to Nash equilibrium in bilinear problems with bandit feedback.
problem Learning dynamics in bilinear saddle-point problems with bandit feedback.
method Uncoupled learning algorithm combining experimental design and FTRL with a tailored regularizer.
result Last-iterate convergence rate of i l d e O ( T − 1 / 4 ) ilde{O}(T^{-1/4}) i l d e O ( T − 1/4 ) in high probability. Paper analyzes convergence rates for multi-agent learning in games.
problem Convergence rates for multi-agent learning in games.
method Characterizes finite-time convergence rates for joint OGD learning on λ λ λ -cocoercive games and develops adaptive algorithms. result Adaptive algorithms achieve same convergence rates as non-adaptive counterparts.
New framework reduces fault tolerance costs in machine learning.
problem Fault tolerance in iterative-convergent machine learning algorithms.
method Developed a general framework to quantify and design strategies for checkpoint-based fault tolerance.
result SCAR reduces iteration cost of partial failures by 78% - 95%.
New method shows last-iterate convergence in constrained min-max optimization.
problem Min-max optimization with constraints.
method Optimistic Multiplicative-Weights Update (OMWU) method.
result Last-iterate convergence to saddle points in constrained min-max optimization.
Mutation improves FTRL convergence in zero-sum games.
problem Lack of last-iterate convergence in FTRL variants.
method Introduced mutation to perturb action probabilities in FTRL.
result M-FTRL converges to Nash equilibria under full-information feedback.
New convergence rates for shuffling gradient methods without strong convexity.
problem Theoretical gap between shuffling gradient methods' empirical success and established convergence rates.
method Proved last-iterate convergence rates for shuffling gradient methods using function value gap.
result First last-iterate convergence rates for shuffling gradient methods without strong convexity.
Two accelerated extragradient methods converge at O ( 1 / k ) O(1/k) O ( 1/ k ) rate for co-hypomonotone inclusions.
problem Solving co-hypomonotone inclusions with sum of Lipschitz and multivalued operators.
method Developed two Nesterov's accelerated extragradient methods for co-hypomonotone inclusions.
result Achieve O ( 1 / k ) \mathcal{O}(1/k) O ( 1/ k ) last-iterate convergence rates on the residual norm. This paper advances extragradient methods for solving inclusions under co-hypomonotonicity.
problem Solving inclusions with non-Lipschitz mappings.
method Unified and generalized extragradient methods under co-hypomonotonicity.
result Achieved O ( 1 / k ) \mathcal{O}(1/k) O ( 1/ k ) convergence rates for residual norms. Improved shuffling gradient methods converge faster for nonsmooth convex optimization.
problem Improving convergence rates for nonsmooth convex optimization problems.
method Analysis of shuffling gradient methods, focusing on Random Reshuffle and Single Shuffle strategies.
result Shuffling gradient methods, particularly Random Reshuffle and Single Shuffle, converge faster than Proximal Gradient Descent for nonsmooth convex optimization.
New study shows faster convergence of SGD and Kaczmarz methods.
problem Improving convergence rates of iterative linear system solvers.
method Last-iterate convergence analysis of SGD with greedy step size over smooth quadratics.
result The t t t -th iterate attains an O ( 1 / t 3 / 4 ) O(1/t^{3/4}) O ( 1/ t 3/4 ) convergence rate. Gradient methods converge exponentially in concave network games.
problem Finding Nash equilibria in concave network zero-sum games.
method Gradient Ascent and Optimistic Gradient Ascent analyses.
result Exponential convergence rates in various game settings.
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.
New algorithm for solving minimax problems over distributions converges to Nash equilibrium.
problem Solving minimax problems over probability distributions.
method Symmetric Mean-field Langevin Dynamics (MFL-AG and MFL-ABR) with weighted averaging and best response dynamics.
result Converges to mixed Nash equilibrium with average-iterate and last-iterate convergence.
Non-affine aggregation rules cannot preserve monotonicity in convex learning.
problem Designing non-affine aggregation rules that maintain monotonicity in convex learning.
method Proving that monotonicity of aggregated gradients is preserved only if the aggregation rule is positively affine.
result Non-affine aggregation prevents steady convergence and substantially degrades algorithmic stability.
Convex message passing algorithms converge to a fixed point.
problem Understanding convergence properties of convex message passing methods.
method Proving convergence of coordinate descent applied to piecewise-affine convex objectives, and showing this applies to various message passing methods.
result The iterates converge to a fixed point of the method, and the algorithm terminates in a known number of iterations.
A new method for RLHF using proximal point Nash learning.
problem Capturing real human preferences in RLHF.
method Proximal point Nash learning, embedding self-play updates into a proximal point framework.
result High-probability last-iterate convergence for the combined method.
WSqD extends learning rate schedules for large model training without fixed horizons.
problem Fixed learning rate schedules limit training horizon extension.
method WSqD replaces constant stable phase with a shifted inverse-square-root base, retaining linear cooldown.
result WSqD achieves minimax-optimal convergence rate and horizon-independence.
Paper generalizes extragradient methods for solving equations and inclusions with improved convergence rates.
problem Solving equations and inclusions using extragradient methods.
method Unified and generalized extragradient methods for a broader class of algorithms, analyzing sublinear convergence rates.
result Unified and improved convergence results for various extragradient variants.
Develops new algorithms for solving root-finding problems in large-scale settings.
problem Solving nonlinear equations in large-scale settings.
method Randomized block-coordinate optimistic gradient algorithms.
result Achieves convergence rates of O ( 1 / k ) \mathcal{O}(1/k) O ( 1/ k ) and O ( 1 / k 2 ) \mathcal{O}(1/k^2) O ( 1/ k 2 ) for root-finding problems. 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…
This paper studies iteration convergence of Kronecker graphical lasso (KGLasso) algorithms for estimating the covariance of an i.i.d. Gaussian random sample under a sparse Kronecker-product covariance model and MSE convergence rates. The KGlasso model, originally called the transposable regularized covariance model by …
Study on curvature image iterations converging to solutions of Minkowski problems.
problem Solving L p L_p L p Minkowski problems with prescribed data. method Iterations of curvature image operators Λ p φ Λ_p^{\varphi} Λ p φ applied to convex bodies. result Iterations of curvature image operators converge to fixed points under certain conditions.
New bounds on SGD's final iterate convergence rate in constant dimension.
problem Characterize the convergence rate of SGD's final iterate in constant dimension.
method Proved lower bounds of Ω ( log d / T ) Ω(\log d/\sqrt{T}) Ω ( log d / T ) and Ω ( log d / T ) Ω(\log d/T) Ω ( log d / T ) for non-smooth Lipschitz convex and strongly convex functions respectively. result First general dimension dependent lower bound on SGD's final iterate convergence rate.
Gradient descent finds the shortest path to global optima in overparameterized models.
problem Training overparameterized nonlinear models with gradient descent.
method Gradient descent, martingale techniques.
result Gradient descent converges geometrically to a global optimum near the initial point.
Improves training GANs by escaping limit cycles.
problem Limit cycling behavior in training GANs.
method Predictive Centripetal Acceleration Algorithm (PCAA) combined with Adam.
result PCAA improves convergence rates and effectively trains GANs.
The paper extends gap theorems for Bach-flat 4-manifolds.
problem Proving gap theorems for specific Bach-flat 4-manifolds.
method Iteration argument and convergence theory of Bach-flat metrics.
result Conformally invariant gap theorems for ( C P 2 , g F S ) (\mathbb{CP}^2, g_{FS}) ( CP 2 , g F S ) and ( S 2 i m e s S 2 , g p r o d ) (\mathbb{S}^2 imes\mathbb{S}^2,g_{prod}) ( S 2 im es S 2 , g p r o d ) . This work analyzes and improves stochastic gradient methods for GAN training.
problem Understanding the training dynamics of GANs, particularly their convergence.
method Continuous-time analysis using differential equations, focusing on simGD and its variants.
result The methods converge under different assumptions, providing new insights into GAN training.
Mirror flow optimizes separable data problems, converging to a maximum margin classifier.
problem Optimizing classification problems with separable data using mirror flow.
method Examine mirror flow on linearly separable classification problems, focusing on the horizon function of the mirror potential.
result Mirror flow converges to a maximum margin classifier for separable data under certain conditions.
A new algorithm LONR learns without terminal states or perfect recall.
problem Learning in settings without terminal states or perfect recall.
method Local No-Regret Learning (LONR) using Q-learning-like updates.
result LONR achieves last iterate convergence in challenging settings.
The sum-product or belief propagation (BP) algorithm is a widely used message-passing technique for computing approximate marginals in graphical models. We introduce a new technique, called stochastic orthogonal series message-passing (SOSMP), for computing the BP fixed point in models with continuous random variables.…
Optimizes convergence rate of stochastic proximal algorithms for composite convex problems.
problem Solving composite convex optimization problems with composite regularizers.
method Analyzed proximal stochastic gradient method and randomized incremental proximal method under relaxed variance assumptions.
result Proves O ( 1 / T ) O(1/\sqrt{T}) O ( 1/ T ) convergence rate for last iterate of both algorithms under componentwise convexity and smoothness. This work studies the contraction coefficients of Schrödinger bridge problems in linear systems.
problem Optimally controlling the evolution of a system's state density over time.
method Analyzes and improves the convergence rates of dynamic Schrödinger systems via geometric and control-theoretic interpretations.
result New insights into improving computation of worst-case contraction coefficients by preconditioning.
New framework uses PDE for no-regret generative modeling.
problem Developing efficient generative models for complex distributions.
method Iterative refinement of Brenier maps using mirror gradient descent.
result Converges to optimal Brenier map under various step-size schedules.
Last iterate of Extragradient algorithm converges slower than averaged iterates in saddle point problems.
problem Smooth convex-concave saddle point problems
method Analysis of Extragradient (EG) algorithm convergence rates
result The last iterate of EG converges at a rate of O(1/√T), compared to O(1/T) for averaged iterates
Proposes a method to reduce parallel complexity of MLMC in SGD.
problem Poor scalability of MLMC in SGD on parallel platforms.
method Proposes a delayed MLMC gradient estimator to reduce parallel complexity.
result Proves reduction in average parallel complexity per iteration at the cost of slightly worse convergence rate.
New algorithm robustly optimizes data streams with heavy-tailed or infinite variance samples.
problem Optimizing data streams with heavy-tailed or infinite variance samples.
method Gradient quantile clipping for SGD, leveraging Markov chain connections.
result Algorithm converges to a concentrated distribution with high probability bounds.
SGD converges to an invariant distribution with sub-Gaussian or sub-exponential properties.
problem Optimizing smooth and strongly convex objectives using SGD.
method Analysis through Markov chains, focusing on convergence and concentration properties.
result SGD iterates and their invariant limit distribution inherit sub-Gaussian or sub-exponential concentration properties.
A new algorithm solves semidefinite programs using Langevin diffusion.
problem Optimizing semidefinite programs with diagonal constraints.
method Langevin diffusion on a product manifold of spheres.
result Langevin algorithm achieves ε accuracy in Ω(ε^-5) iterations.
Riemannian algorithms converge at Euclidean rates for geodesically convex-concave problems.
problem Min-max optimization on Riemannian manifolds.
method RCEG method and RGDA for geodesically strongly-convex-concave problems.
result RCEG achieves linear convergence rate in geodesically strongly-convex-concave cases.
New method implicitly regularizes high-dimensional linear regression using gradient descent.
problem Sparse vector estimation in high-dimensional linear regression.
method Gradient descent on residual sum of squares with early stopping under overparameterization.
result Gradient descent implicitly leads to nearly sparse optimal solutions without explicit penalties.
New algorithm achieves sublinear regret in CMDPs without error cancellations.
problem Safety constraints in reinforcement learning with error cancellations.
method Model-based primal-dual algorithm for CMDPs with multiple constraints.
result Achieves sublinear regret without error cancellations.
In distributed ML applications, shared parameters are usually replicated among computing nodes to minimize network overhead. Therefore, proper consistency model must be carefully chosen to ensure algorithm's correctness and provide high throughput. Existing consistency models used in general-purpose databases and moder…