Applying Adaptive Gradient Descent to solve matrix factorization

Based on the method of FGD, we apply the method of adaptive gradient descent which uses different step length at different epoch. Adaptive gradient descent performs much better than FGD in the tests and keeps the guarantee of convergence speed at the same time.

Computer Science Mathematics Engineering

View graph Source

Paper

Similar papers

Gradient descent in matrix factorization: Understanding large initializationHengchao Chen, Xin Chen, Mohamad Elmasri et al.2023 · Conference on Uncertainty in Artificial Intelligence · 2 citations In Library

Accelerated Scaled Gradient Descent Algorithm For Low-Rank Matrix FactorizationXiaojing Jia, Xiangchu Feng, H.Y. Wu2024 · 2024 5th International Conference on Computer Engineering and Intelligent Control (ICCEIC) · 0 citations In Library

Convergence of Alternating Gradient Descent for Matrix FactorizationRachel Ward, Tamara G. Kolda2023 · arXiv (Cornell University) · 14 citations In Library

Fractional Gradient Descent With Matrix Stepsizes for Non-Convex OptimizationAlokendu Mazumder, Keshav Vyas, Punit Rathore2025 · IEEE Transactions on Neural Networks and Learning Systems · 0 citations In Library

An Efficient Parallel Stochastic Gradient Descent for Matrix Factorization On GPUST. Y. Xing, Bin Wu, Bai Wang2020 · International Conference on Data Science in Cyberspace · 0 citations In Library

Accelerating Ill-Conditioned Low-Rank Matrix Estimation via Scaled Gradient DescentTian Tong, Cong Ma, Yuejie Chi2020 · Journal of machine learning research · 155 citations In Library

Improved Gradient Descent Optimization Using Adaptive Step-SizingDivya Sharma, Vikram Sethi2023 In Library

AdaGDA: Faster Adaptive Gradient Descent Ascent Methods for Minimax OptimizationFeihu Huang, Heng Huang2021 · International Conference on Artificial Intelligence and Statistics · 25 citations In Library

Functional Gradient Descent with Adaptive RepresentationsDaniel Csillag, R. Schuller, Pedro Dall'Antonia et al.2026 · 0 citations In Library

Exact Linear Convergence Rate Analysis for Low-Rank Symmetric Matrix Completion via Gradient DescentTrung Vu, R. Raich2021 · IEEE International Conference on Acoustics, Speech, and Signal Processing · 10 citations In Library

Patents near this paper

Matrix-Factorization Based Gradient CompressionUS 11,182,457 · INTERNATIONAL BUSINESS MACHINES CORPORATION · 2021 Fast Distributed Nonnegative Matrix Factorization and Completion for Big Data AnalyticsUS 10,304,008 · NEC LABORATORIES AMERICA, INC. · 2019 ADAPTIVE FILTERING USING FAST AFFINE PROJECTION ADAPTATIONUS 7,436,880 · NATIONAL RESEARCH COUNCIL OF CANADA · 2008 SYSTEMS AND METHODS USING ANGLE-BASED STOCHASTIC GRADIENT DESCENTUS 11,347,997 · THE FLORIDA INTERNATIONAL UNIVERSITY BOARD OF TRUSTEES · 2022 GAUSSIAN RANKING USING MATRIX FACTORIZATIONUS 10,180,968 · NETFLIX, INC. · 2019

Source of record

arXiv