archive
Every paper Pith has read. Search by title, abstract, or pith.
999 papers in math.OC · page 2
-
Spectral clipping achieves optimal rate for heavy-tailed SGD
Gradient Clipping Beyond Vector Norms: A Spectral Approach for Matrix-Valued Parameters
-
Proximal limited-memory quasi-Newton converges globally for nonconvex problems
Proximal Limited-Memory Quasi-Newton Methods for Nonsmooth Nonconvex Optimization
-
Neumann boundary data observe waves on gas giant metrics
Boundary observability for gas giant metrics
-
Strong duality holds for weakly communicating average-reward CMDPs
Learning Weakly Communicating Average-Reward CMDPs: Strong Duality and Improved Regret
-
HS-Jacobian lets Adam train neural nets with linear constraints
Efficient and provably convergent end-to-end training of deep neural networks with linear constraints
-
Barrier smoothing yields O(K^{-2/3}) stationarity for constrained bilevel opt
A Barrier-Metric First-Order Method for Linearly Constrained Bilevel Optimization
-
DNN relaxation exact for random quadratics with high probability
Exactness of the DNN Relaxation for Random Standard Quadratic Programs
-
Schrödinger bridge solves sub-Riemannian optimal transport
From Schrodinger Bridge to Optimal Transport over Sub-Riemannian Manifolds
-
Certificate establishes exact worst-case rate for gradient descent when N >= 3
The Grimmer-Shu-Wang Certificate and the Drori-Teboulle Minimax Nonnegative Constant-Stepsize Bound for N >= 3
-
Reputation learning closes loop on Byzantine consensus
Byzantine-Resilient Consensus via Active Reputation Learning
-
Gauss-Newton whitens errors to outperform Newton
Error whitening: Why Gauss-Newton outperforms Newton
-
Quotient symmetry fixes distributional average-reward RL
Quotient-Categorical Representations for Bellman-Compatible Average-Reward Distributional Reinforcement Learning
-
Mirror descent computes exact barycenters for discrete and continuous measures
A Unified Approach for Computing Wasserstein Barycenters of Discrete and Continuous Measures
-
Random spectra match Muon on GPT-2 training
Muon is Not That Special: Random or Inverted Spectra Work Just as Well
-
Separable estimators tighten relaxations beyond McCormick
Relaxation via Separable Estimators: Arithmetic and Implementation
-
Projected JSR can be strictly smaller than γ for deflated Q-VI
Switching-Geometry Analysis of Deflated Q-Value Iteration
-
Single network solves optimal transport via proximal fixed points
Fixed-Point Neural Optimal Transport without Implicit Differentiation
-
Gradient descent reaches only global minima in wide shallow nets
On the global convergence of gradient descent for wide shallow models with bounded nonlinearities
-
Decentralized MPC with safe sets guarantees multi-agent collision avoidance
Decentralized Contingency MPC based on Safe Sets for Nonlinear Multi-agent Collision Avoidance
-
Exponential bound proven for LCP sufficient-matrix handicaps
Handicap reduction for linear complementarity problems
-
Natural policy gradient equals smoothed policy iteration
Natural Policy Gradient as Doubly Smoothed Policy Iteration: A Bellman-Operator Framework
-
Closed-form spectral formulas estimate density ratios from moments
A Spectral Framework for Closed-Form Relative Density Estimation
-
New moves link all incomplete tournament schedules
Novel neighborhood structures for incomplete round robin sports tournaments
-
Frank-Wolfe lower bound matches upper bound on p-uniformly convex sets
Curvature-Dependent Lower Bounds for Frank-Wolfe
-
Riemannian L-BFGS handles Euclidean bounds on manifolds
A Riemannian quasi-Newton algorithm for optimization with Euclidean bounds
-
LP methods give exact set tolerances for MST
Computation of Set Tolerances with Applications to the Minimum Spanning Tree Problem
-
8/3 approximation for matroid-constrained randomized vertex-cover interdiction
Randomized Max-Vertex-Cover Interdiction with Matroid Constraints
-
Backstepping observer stabilizes error in blood flow cascade models
Observer Design for a Class of ODE -- Continuum-PDE Cascade Systems Inspired by a Control-Theoretic Model of Large-Scale Arterial Networks of Blood Flow
-
Bound certifies any learned controller for unknown linear systems
A PAC-Bayes Approach for Controlling Unknown Linear Discrete-time Systems
-
Youla-Kucera adds channels for cascaded MPC and offset-free control
Hierarchical 2-degree-of-freedom control combining Youla-Kucera parameterization and model predictive control
-
LLM writes branching rules that speed up MILP solvers
LLM4Branch: Large Language Model for Discovering Efficient Branching Policies of Integer Programs
-
Attention fuses LEO measurements for spectrum cartography
Learning-Based Spectrum Cartography in Low Earth Orbit Satellite Networks: An Overview
-
PowerStep matches Adam on Transformers with half the optimizer memory
PowerStep: Memory-Efficient Adaptive Optimization via $\ell_p$-Norm Steepest Descent
-
Signature method gives sublinear regret for path-dependent bandits
Signature Approach for Contextual Bandits with Nonlinear and Path-dependent Rewards
-
BROS matches exact bilevel convergence while cutting peak memory by up to 45%
BROS: Bias-Corrected Randomized Subspaces for Memory-Efficient Single-Loop Bilevel Optimization
-
Randomized subspaces match exact bilevel convergence rate
BROS: Bias-Corrected Randomized Subspaces for Memory-Efficient Single-Loop Bilevel Optimization
-
XP algorithms and W[1]-hardness classify stationarity testing for PA functions in fixed d
Parameterized Complexity of Stationarity Testing for Piecewise-Affine Functions and Shallow CNN Losses
-
Phased algorithm achieves d sqrt(T) regret for sparse linear bandits
Learning to Sparsify Stochastic Linear Bandits
-
Transformation stabilizes ODE-wave cascade with boundary disturbances
Stabilization for a Cascaded ODE-Wave Equation with Boundary Nonlinear Disturbances
-
Same-optimizer solutions form connected sets in wide ReLU nets
Optimizer-Induced Mode Connectivity: From AdamW to Muon
-
Chebyshev center selects PINN update directions
Chebyshev Center-Based Direction Selection for Multi-Objective Optimization and Training PINNs
-
Order-gap measure gives stopping rule for adaptive learning
Consolidation-Expansion Operator Mechanics:A Unified Framework for Adaptive Learning
-
Order-gap tracks distance to settled state in learning systems
Consolidation-Expansion Operator Mechanics:A Unified Framework for Adaptive Learning
-
Backward LPs yield optimal recommendations to strategic agents
Action Recommendations for Sequentially Rational Strategic Agents
-
Vector measurements speed up Bayesian optimization
Bayesian Optimization with Structured Measurements: A Vector-Valued RKHS Framework
-
Probabilistic sets let Gaussian processes safely explore nonlinear systems
Safe Exploration for Nonlinear Processes Using Online Gaussian Process Learning
-
Power law model splits Muon and SignSGD into three phases
Phases of Muon: When Muon Eclipses SignSGD
-
Certificates isolate Koopman regression failures by layer
Diagnostic Certificates of Data Quality and Regression Identifiability for Koopman Identification
-
Mean-field SVGD converges in L2 at explicit polynomial rates
Quantitative Local Convergence of Mean-Field Stein Variational Gradient Flow
-
Mobile multiplicative control steers quasilinear parabolic equations to rest
Controllability of quasilinear parabolic equations under multiplicative mobile controls