About this Research Topic
Deep learning is a new branch of black-box state-of-the-art machine learning algorithms that have been proven to be a powerful feature extraction tool in computer vision. However, there are some limitations of deep learning about: (1) mechanism for learning abstraction through large numbers of training examples, e.g. convolutional neural networks will face exponential inefficiencies, (2) limited capacity for transfer knowledge, (3) insufficiently transparent, i.e. opacity of black-box deep neural networks, (4) not well integrated with prior knowledge, (5) working well as a universal approximator but hard to be fully trusted. Surprisingly, little work has been conducted in relation to these limitations of deep learning. Bridging these gaps could be expected to contribute to greatly impact a variety of real-world engineering applications such as the safety and reliability of fully automated driving, medical applications such as digital pathology and radiology, and material science applications.
We welcome all types of articles. Potential topics include, but are not limited to the following:
• Use of deep learning for explainable AI,
• Perspectives on deep learning in small numbers of training examples,
• Transferring knowledge for deep learning,
• Approximation capability by deep learning with well trusted,
• Improvements of opacity of black-box deep neural networks,
• Various applications in deep learning with reduced limitations, e.g. Learning, Cognitive Science, Neuropsychology, Systems Biology, Information Theory, Digital Pathology, Radiology, Diabetology.
Keywords: Deep Learning, Transparency of deep neural networks, Explainable AI, Effective use of Deep Learning, Engineering and medical applications, Learning, Cognitive Science, Neuropsychology, Systems Biology, Information Theory, Digital Pathology, Radiology, Diabetology
Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.