In today's healthcare system, artificial intelligence (AI), wearable devices, and access to high-speed internet have dramatically changed the way healthcare services are delivered. Diagnostic markers promise to be a big success as they provide biological parameters that aid the diagnosis of diseases. Diagnostic markers can be used to identify individuals with a particular type of disease or condition, or to detect the presence of the condition. Biomarkers that identify disease subtypes thus often play crucial roles in prognosticating and predicting disease outcomes in the case of diagnostic classification. It often takes a lot of time and money to conduct traditional research on diagnostic markers. It is difficult to synthesize large amounts of data since many research objects include dozens of samples. Therefore, some researchers have used machine learning and deep learning models for diagnostic markers. Since deep learning models can extract potential features from the input data automatically, researchers may prefer these models since handcrafted features are not needed to build the model.

However, the deep learning models are sensitive to their architecture and hyperparameters. The design of deep learning architectures is a difficult and time-consuming task that also requires specialized knowledge. Hyperparameters are all variables that are defined manually at the beginning of training. For better performance, automated approaches are required which can effectively select the architecture and hyperparameters of the deep learning models.

This Special Issue aims to optimize the architecture and hyperparameters of deep learning models in order to achieve better performance of diagnostic markers. The architecture and initial parameters of deep learning models can be efficiently determined using metaheuristic techniques. However, the development of metaheuristics-based optimized deep learning models for diagnostic biomarkers is an area that needs to be explored further. We welcome original research and review articles.

Potential topics include but are not limited to the following:

• Optimized deep learning models for diagnostic markers
• Hardware implementation of optimized models for diagnostic markers
• Complexity analyses of optimized deep learning models for diagnostic markers
• Optimized deep learning models for Medical Internet of Things (MIoT) devices
• Optimized pre-trained models for diagnostic markers’ applications
• Explainable optimized deep learning models for diagnostic markers
• Optimized deep federated learning models for diagnostic markers
• Secure and verifiable deployment of deep learning models for diagnostic markers
• Blockchain-enabled deployment of deep learning models for diagnostic markers