Abstract:Because of the good performance of Convolutional Neural Network (CNN), it has been extensively used in many fields, such as image, speech, text, etc. However, it is easily affected by hyperparameters. How to effectively configure hyperparameters at a reasonable time to improve the performance of convolutional neural networks has always been a complex problem. To solve this problem, this paper proposes a method to automatically optimize CNN hyperparameters based on the local autonomous competitive harmony search (LACHS) algorithm. To avoid the influence of complicated parameter adjustment of local autonomous competitive harmony search algorithm on its performance, a parameter dynamic adjustment strategy is adopted, which makes the pitch adjustment probability (PAR) and step factor BW dynamically adjust according to the actual situation. To strengthen the fine search of neighborhood space and reduce the possibility of falling into local optima for a long time, an autonomous decision-making search strategy based on the optimal state is designed. To help the algorithm jump out of the local fitting situation, this paper proposes a local competition mechanism to make the new sound competes with the worst harmonic progression of local selection. In addition, an evaluation function is proposed, which integrates the training times and recognition accuracy. To achieve the purpose of saving the calculation cost without affecting the search result, it makes the training time for each model depending on the learning rate and batch size. In order to prove the feasibility of LACHS algorithm in configuring CNN superparameters, the classification of the Fashion-MNIST dataset and CIFAR10 dataset is tested. The comparison is made between CNN based on empirical configuration and CNN based on classical algorithms to optimize hyperparameters automatically. The results show that the performance of CNN based on the local autonomous competitive harmony search algorithm has been improved effectively, so this algorithm has certain advantages in hyperparametric optimization. In addition, this paper applies the local autonomous competitive harmony search algorithm to expression recognition. Experiments show that the performance of CNN optimized based on the local autonomous competitive harmony search algorithm is better than that of the same type of artificially designed CNN. Therefore, the method proposed in this paper is feasible in practical application.

Application of Metaheuristic Algorithms for Determining the Structure of a Convolutional Neural Network with a Small Dataset

Hyperparameters Optimization of Convolutional Neural Network Based on Local Autonomous Competition Harmony Search Algorithm

Metaheuristic Algorithms for Convolution Neural Network

Convolution Neural Network Hyperparameter Optimization Using Simplified Swarm Optimization

Small sample classification of hyperspectral image using model-agnostic meta-learning algorithm and convolutional neural network

Hyperspectral Image Classification Promotion Using Dynamic Convolution Based on Structural Re-Parameterization

A Pixel Cluster Cnn And Spectral-Spatial Fusion Algorithm For Hyperspectral Image Classification With Small-Size Training Samples

Optimization of Convolutional Neural Network Hyperparameter for Medical Image Diagnosis using Metaheuristic Algorithms: A short Recent Review (2019-2022)

Surrogate-Assisted Hybrid-Model Estimation of Distribution Algorithm for Mixed-Variable Hyperparameters Optimization in Convolutional Neural Networks

Convolutional Neural Networks Based Hyperspectral Image Classification Method with Adaptive Kernels

Genetic Algorithm Based Deep Learning Neural Network Structure and Hyperparameter Optimization

Structural Analysis and Optimization of Convolutional Neural Networks with a Small Sample Size

Hyperparameter Recommendation Integrated With Convolutional Neural Network

Fuzzy Cerebellar Model Articulation Controller Network Optimization Via Self-Adaptive Global Best Harmony Search Algorithm

A Hybrid Training Method of Convolution Neural Networks Using Adaptive Cooperative Particle Swarm Optimiser.

An Investigation of a Multidimensional CNN Combined with an Attention Mechanism Model to Resolve Small-Sample Problems in Hyperspectral Image Classification

Optimal Parameter Selection In Hyperspectral Classification Based On Convolutional Neural Network

Efficient Hyperparameter Optimization in Deep Learning Using a Variable Length Genetic Algorithm

Divide-And-Conquer Dual-Architecture Convolutional Neural Network For Classification Of Hyperspectral Images

Optimization of Convolutional Neural Network using Microcanonical Annealing Algorithm