Kwangjun Ahn
21 papers · 2018–2025 · 6 venues · across top CS/AI venues
Authors are matched by name, so researchers who share a name ("Yang Liu", "Wei Wang") appear as a single profile and their paper counts, venues and achievements are combined. How identity is handled.
Achievements
Jump to papers ↓+7 more ↓ Show less ↑
🗺️ Taxonomy Completionist (36) 🏃 Academic Marathon · Sprint 🌍 Conference Polyglot · Silver 🌉 Interdisciplinary Bridge · Bronze 🐣 Hot Topic Early Bird
👑
Triple Crown
⚡
Prolific Year · Productive
🔥
Unstoppable · Bronze
❓
The Questioner
💎
Century Club · Bronze
🌈
Renaissance Researcher
(5)
🗃️
Keyword Collector
(100)
Top co-authors
Research topics
Machine Learning
(19)
Optimization & Theory · Machine Learning
(18)
Mathematics & Optimization
(8)
Optimization
(8)
Deep Learning
(7)
Core Methods
(4)
Mathematics
(3)
Optimization & Theory · Deep Learning
(3)
Techniques
(3)
Application Areas
(2)
Architectures
(2)
Learning Types
(2)
Artificial Intelligence
(1)
Data Science & Analytics
(1)
Natural Language Processing
(1)
Keywords
non-convex optimization
(4)
gradient descent
(3)
sharpness-aware minimization
(3)
stochastic gradient descent
(3)
convex optimization
(3)
mirror descent
(2)
maximum-margin solution
(2)
edge of stability
(2)
overparameterized model
(2)
training dynamics
(2)
log-concave distribution
(2)
convergence rate
(2)
langevin dynamics
(2)
markov chain monte carlo
(2)
adam optimizer
(2)
sampling algorithm
(2)
convergence analysis
(1)
global convergence
(1)
in-context learning
(1)
optimal transport
(1)
Papers
Does SGD really happen in tiny subspaces?
ICLR 2025
General framework for online-to-nonconvex conversion: Schedule-free SGD is also effective for nonconvex optimization
ICML 2025
The Belief State Transformer
ICLR 2025
Adam with model exponential moving average is effective for nonconvex optimization
NeurIPS 2024
How to Escape Sharp Minima with Random Perturbations
ICML 2024
Linear attention is (maybe) all you need (to understand Transformer optimization)
ICLR 2024
Understanding Adam Optimizer via Online Learning of Updates: Adam is FTRL in Disguise
ICML 2024
A Unified Approach to Controlling Implicit Regularization via Mirror Descent
JMLR 2023
Learning threshold neurons via edge of stability
NeurIPS 2023
Model Predictive Control via On-Policy Imitation Learning
L4DC 2023
The Crucial Role of Normalization in Sharpness-Aware Minimization
NeurIPS 2023
Transformers learn to implement preconditioned gradient descent for in-context learning
NeurIPS 2023
Agnostic Learnability of Halfspaces via Logistic Loss
ICML 2022
Mirror Descent Maximizes Generalized Margin and Can Be Implemented Efficiently
NeurIPS 2022
Reproducibility in Optimization: Theoretical Framework and Limits
NeurIPS 2022
Understanding the unstable convergence of gradient descent
ICML 2022
Efficient constrained sampling via the mirror-Langevin algorithm
NeurIPS 2021
Optimal dimension dependence of the Metropolis-Adjusted Langevin Algorithm
COLT 2021
From Nesterov’s Estimate Sequence to Riemannian Acceleration
COLT 2020
SGD with shuffling: optimal rates without component convexity and large epoch requirements
NeurIPS 2020
Binary Rating Estimation with Graph Side Information
NeurIPS 2018