#convex optimization
19 papers match
Towards joint scaling laws with optimal batch size schedules
Jiaxiang Li, Zhiqi Bu, Shiyun Xu
The paper derives a theoretical relationship between learning rate and batch size schedules using convex optimization, and proposes a closed‑form optimal batch size schedule that i…
Normalized First-Order Methods for Convex (L0, L1)-Smooth Optimization with Inexact Gradients
Evgeniy Kovalev, Fedor Stonyakin
The paper proposes and analyzes convex optimization algorithms that work with a comparison oracle providing inexact normalized gradients for (L0, L1)-smooth problems, establishing…
Entropy-Smooth Convex Optimization Cannot Be Accelerated
Jacob M. Aguirre, Dmitrii M. Ostrovskii
The paper proves that for convex functions that are smooth relative to negative entropy (or von Neumann entropy in the quantum case), no first-order method can achieve an accelerat…
A Unifying Framework for Quasi-Polynomial Optimization of Fixed-degree Polynomials
Martino Bernasconi, Matteo Castiglioni, Andrea Celli +1
The paper presents a method to construct ε‑covers for the joint value sets of constant-degree polynomials over convex domains, enabling quasi‑polynomial time approximation schemes…
Modeling and Control of Deep Sign-Definite Dynamics with Application to Hybrid Powertrain Control
Teruki Kato, Ryotaro Shima, Kenji Kashima
The paper proposes enforcing sign constraints on the Jacobian of deep neural network models to ensure physical properties like monotonicity and positivity, enabling convex model pr…
What's in a Smoothness Constant? Tighter Rates for Local SGD with Bounded Second-order Heterogeneity
Kumar Kshitij Patel, Rustem Islamov, Sebastian U Stich +3
The paper establishes tighter convergence rates for Local SGD (Federated Averaging) on general convex problems under a bounded second‑order heterogeneity assumption, and provides n…
Security evaluation of quantum distance-bounding protocols via semidefinite programming
Kevin Bogner, Aysajan Abidin, Dave Singelee +1
The paper analyzes the security of quantum distance‑bounding protocols by formulating one‑round distance‑fraud and mafia‑fraud attacks as convex optimization problems that can be s…
An adaptive interior-point method with backtracking line search for convex constrained optimization
Fadi Hamad, Oliver Hinder
The paper proposes and analyzes an adaptive interior‑point algorithm that uses a regularized Newton step with backtracking line search on a log‑barrier function to solve convex pro…
A Two-step Krasnosel'skii-Mann Algorithm with Adaptive Momentum and Its Applications to Image Denoising and Matrix Completion
Yongxin He, Jingyuan Li, Yizun Lin +1
The paper introduces a two-step Krasnosel'skii‑Mann algorithm with adaptive momentum for solving convex optimization problems, and demonstrates its effectiveness on image denoising…
Tight Lower Bounds for Binary First-Order Oracles for Convex Optimization
Amitabh Basu, Phillip Kerger, Marco Molinaro
The paper proves that solving mixed-integer convex optimization with bit‑wise first‑order oracles requires a number of bits that grows quadratically with the number of continuous v…
Estimating Distributions with Failure Rate Properties from Noisy Quantile Data
Timothy C. Y. Chan, Ningyuan Chen, Craig Fernandes +1
The paper develops a method to estimate an unknown cumulative distribution function under increasing failure‑rate shape constraints using noisy quantile observations, providing a t…
Closing the Oracle-Complexity Gap in Derivative-Free Convex Optimization: A Near-Quadratic Lower Bound from Exact Function Values
Phillip Kerger
The paper proves a near‑quadratic lower bound on the number of exact function‑value queries needed to minimize a convex Lipschitz function over a Euclidean ball, closing a long‑sta…
Nonasymptotic Analysis of Accelerated Methods With Inexact Oracle Under Absolute Error Bound
Yin Liu, Sam Davanloo Tajbakhsh
The paper derives explicit nonasymptotic convergence bounds for two accelerated first‑order methods applied to smooth convex problems when gradients are accessed with bounded absol…
Faster Newton Methods for Convex and Nonconvex Optimization in Gradient Complexity
Lesi Chen, Chengchang Liu, Luo Luo +1
The paper proposes new second‑order optimization algorithms that reduce the gradient complexity for both convex and nonconvex problems, establishing tighter theoretical bounds than…
Decentralized Model Predictive Control of Connected and Automated Vehicles with Coupled Safety Constraints
Philip Schultheis, Kimia Chavoshi, John Lygeros
The paper proposes decentralized model predictive control methods for coordinating connected and automated vehicles on lane-free highways, using a new technique to turn complex saf…
Lecture Notes: Convex Optimization
Andreas Habring
These lecture notes introduce the theory and algorithms of convex optimization, covering existence results, projected subgradient descent, proximal‑gradient and accelerated gradien…
Sparse Robust Optimal Control in Continuous-Time: A Computationally Viable Approach
Siddhartha Ganguly, Ashwin Aravind, Souvik Das +2
The paper introduces a computationally tractable algorithm that solves sparse robust optimal control problems for continuous-time linear noisy systems by reformulating them as semi…
Optimal Parameter-Free First-Order Methods for Convex Optimization with Unknown Growth and Smoothness
Liwei Jiang, Ke Tang, Zhe Zhang
The paper introduces parameter-free first-order optimization methods that automatically adapt to unknown smoothness and growth properties of convex functions, achieving optimal con…
Exact Solutions to a Class of Constrained Optimal Control Problems via Lossless Convexification for Digital Control
Vaibhav Upadhyay, Siddhartha Ganguly, Debasish Chatterjee
The paper presents a lossless convexification method that transforms certain nonconvex, constrained optimal control problems for linear systems into convex ones and solves them exa…
One search, two signals: results blend meaning (embedding similarity, so papers that never use your words still surface) with keyword matches on titles, abstracts and summaries. Free, no sign-in needed.