At present, the emergence of increasingly complex big data brings more challenges to the current big data analysis technology. Complexity is the fundamental difference between complex big data and traditional big data. It is mainly manifested in four aspects: source diversity, type complexity, structure complexity, and internal pattern complexity. At present, there has been a lot of research progress on the source diversity and the type complexity. However, the structure complexity and internal pattern complexity are the difficulties in the analysis of complex big data, among which the internal pattern complexity is the most widely encountered. Many complex big data sets have complex contents (for example, some image data set contains a variety of scenes or a variety of objects), which requires that the processing methods have robust processing capability for a variety of complex objects.

Many complex big data are greatly affected by external factors (for example, the content of an image changes greatly due to the influence of illumination and occlusion), which requires the processing method to be robust to complex changes. Some complex big data sets have large amounts of data or high feature dimensions, and some applications need real-time processing and have high requirements on the calculation efficiency of massive data. Because of its multi-layer nonlinear structure, the deep learning model has a strong feature learning ability, which provides an effective way to solve the above problems. However, differing from traditional big data learning methods, we still need to comprehensively use all kinds of knowledge and means (including text mining, image processing, complex networks, knowledge transfer, graph neural networks, etc.) to study the internal pattern complexity in complex big data.

Therefore, this Special Issue aims to collate original research and review articles that emphasise the important role of deep learning for complex big data analysis, especially for the analysis of internal pattern complexity. It aims to call for state-of-the-art research in the theory, algorithm, modelling, system, and application of deep learning-based complex big data analysis and to demonstrate the latest efforts of relevant researchers.

Potential topics include but are not limited to the following:

• Deep learning methods for analysis of complex big data with internal pattern complexity
• Deep learning methods for the analysis of complex multi-source time-series data
• Deep learning methods for semantic information extraction from complex image and video data with internal pattern complexity
• Deep learning methods for the analysis of multi-source and multi-dimensional complex big data
• Graph neural network methods for complex network data analysis (complex social network analysis, complex power network analysis, graph correlation analysis, etc.)
• Model acceleration for deep learning of complex big data