Nikita Doikov

12 papers · 2018–2025 · 4 conferences · across top CS/AI conferences

Achievements

🐝 Cross-Pollinator (13) 🐣 Hot Topic Early Bird 🌉 Interdisciplinary Bridge 🧭 Keyword Pioneer 🌍 Conference Polyglot (4) 🏃 Academic Marathon (7) 🏆 Keyword Champion (2) 💎 Century Club (12)

Conferences

ICML (8) AISTATS (2) COLT (1) NIPS (1)

Top co-authors

Martin Jaggi (6) Yurii Nesterov (3) Peter Richtarik (2) El Mahdi Chayti (2) Sebastian U Stich (2) Filip Hanzely (1) Nicolas Flammarion (1) Anastasia Koloskova (1) Bettina Messmer (1) University Edinburgh (1)

Keywords

convex optimization (6) newton method (3) convergence rate (2) second-order optimization (2) superlinear convergence (2) composite optimization (2) stochastic optimization (2) cubic regularization (2) trust region (1) frank-wolfe method (1) first-order algorithm (1) gradient method (1) krylov subspace (1) tensor method (1) convex function (1) cubic newton method (1) lazy hessian (1) arithmetic complexity (1) polynomial preconditioning (1) stochastic interpretation (1)

Papers

Cubic regularized subspace Newton for non-convex optimization AISTATS 2025

Improving Stochastic Cubic Newton with Momentum AISTATS 2025

On-Device Collaborative Language Modeling via a Mixture of Generalists and Specialists ICML 2025

On Convergence of Incremental Gradient for Non-convex Smooth Functions ICML 2024

Spectral Preconditioning for Gradient Methods on Graded Non-convex Functions ICML 2024

Second-Order Optimization with Lazy Hessians ICML 2023

Linearization Algorithms for Fully Composite Optimization COLT 2023

Polynomial Preconditioning for Gradient Methods ICML 2023

Stochastic Subspace Cubic Newton Method ICML 2020

Inexact Tensor Methods with Dynamic Accuracies ICML 2020

Convex optimization based on global lower second-order models NIPS 2020

Randomized Block Cubic Newton Method ICML 2018