.png){kind=icon}
Return
A stochastic recursive gradient algorithm with inertial extrapolation for non-convex problems and machine learning
.png){kind=icon}
OA
AI
DOI：10.1007/s13042-024-02524-6.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
In recent years, the inertial extrapolation step has gained significant attention due to its capacity to expedite algorithm convergence. This technology has found widespread application across various algorithms. However, within the domain of machine learning, the utilization of extrapolation technology has yielded limited results. Therefore, we apply it to stochastic optimization algorithms to address non-convex and machine learning problems. By integrating the inertial extrapolation step and the modified Barzilai-Borwein (BB) technique into the SARAH framework, we propose an inertial stochastic recurrence gradient method. This method incorporates both the inertial extrapolation step and the improved BB technique. Through theoretical analysis presented in this paper, we demonstrate that the algorithm converges to a global optimum and analyze the linear convergence rate of the non-convex ((lambda) over tilde -gradient-dominated) objective functions. The numerical results obtained from evaluating three widely utilized machine learning problems clearly illustrate the superior performance and practical feasibility of the proposed algorithm.
Keywords:
Extrapolation
Algorithm
Non-convex
Global optimum
Convergence
Numerical results
Journal
IF:
2.7
Papers:
3.2K
Citations:
5.6K
Organization
Cited Papers
ShareSaveShareSaveShareSaveShareSaveAccommodating Stealth Assessment in Serious Games: Towards Developing a Generic Tool.png)
2018 10th International Conference on Virtual Worlds and Games for Serious Applications (VS-Games)
OAAI
2018 10th International Conference on Virtual Worlds and Games for Serious Applications (VS-Games)IF0
OAAIShareSaveShareSaveShareSave
View more