article · Journal of Ambient Intelligence and Humanized Computing
Tuning the hyperparameters of convolutional neural networks is critical for performance, especially within dynamic environments where conditions change over time. A reinforcement learning algorithm called the RL-based optimization algorithm, or ROA, uses Q-learning to automate and improve the hyperparameter selection process for these neural networks. The method was evaluated using two standard benchmark image datasets, namely MNIST and CIFAR-10. Across experimental tests, networks tuned with this reinforcement learning approach consistently outperformed networks operating without hyperparameter optimisation. On the MNIST dataset, the optimised network achieved an accuracy of 98.97 percent across five learning epochs, compared to 97.62 percent for the unoptimised baseline. On the CIFAR-10 dataset, it reached 73.40 percent accuracy over ten epochs, improving upon the unoptimised baseline of 71.73 percent.
Designing effective neural networks requires finding the best internal configuration settings, which can be difficult when conditions change dynamically. Automating hyperparameter tuning with reinforcement learning improves classification accuracy across relatively few training cycles. This offers a more systematic way to configure machine learning models without relying entirely on manual adjustments.
This technique could assist software engineers and data science teams seeking automated methods to configure computer vision models. Because the validation was restricted to standard image classification benchmarks such as MNIST and CIFAR-10, the work remains at an early-stage research level. Demonstration on commercial, production-grade datasets and dynamic real-world environments would be required prior to software integration.
AI-generated from the published abstract. Always read the original work before citing.
Abstract Many real-world applications necessitate optimization in dynamic situations, where the difficulty is to locate and follow the optima of a time-dependent objective function. To solve dynamic optimization problems (DOPs), many evolutionary techniques have been created. However, more efficient solutions are still required. Recently, a new intriguing trend in dealing with optimization in dynamic environments has developed, with new reinforcement learning (RL) algorithms predicted to breathe fresh life into the DOPs community. In this paper, a new Q-learning RL-based optimization algorithm (ROA) for CNN hyperparameter optimization is proposed. Two datasets were used to test the proposed RL model (MNIST dataset, and CIFAR-10 dataset). Due to the use of RL for hyperparameter optimization, very competitive results and good performance were produced. From the experimental results, it is observed that the CNN optimized by ROA has higher accuracy than CNN without optimization. When using the MNIST dataset, it is shown that the accuracy of the CNN optimized by ROA when learning 5 epoch is 98.97%, which is greater than the 97.62% of the CNN without optimization. When using the CIFAR-10 dataset, it is shown that the accuracy of the CNN optimized by ROA when learning 10 epoch is 73.40 percent, which is greater than 71.73% of the CNN without optimization.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.1007/s12652-022-03788-y
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
