Enhancing Neural Network Performance Through Hybrid Optimization Methods: a Comparative Study

EasyChair Preprint 15809

Abstract

This paper explores the enhancement of artificial neural network (ANN) performance through the combination of traditional and modern optimization methods. The main goal is to assess hybrid approaches that incorporate Particle Swarm Optimization (PSO) and conventional gradient-based methods to improve the performance of deep learning models in handling complex and noisy data. Through a comparative analysis in various applications such as image recognition and natural language processing (NLP), the results show that these hybrid methods significantly outperform single-algorithm approaches. This paper presents experimental results alongside detailed analyses and computational complexity assessments of these algorithms.

Keyphrases: Algorithms, PSO, complexity, machine learning

@booklet{EasyChair:15809,
  author    = {Isabel Cheng and James Kung and H Kung and Rashed Ali and Mohammad Azil},
  title     = {Enhancing Neural Network Performance Through Hybrid Optimization Methods: a Comparative Study},
  howpublished = {EasyChair Preprint 15809},
  year      = {EasyChair, 2025}}