查看文章

Over the past 10 years, deep learning has attracted a lot of attention, and many exciting results have been achieved in various areas, such as speech recognition, computer vision, handwriting recognition, machine translation, and natural language understanding. Rather surprisingly, the performance of machines has even surpassed humans’ in some specific areas. The fast development of deep learning has already started impacting people’s lives; however, challenges still exist. In particular, the theory of successful deep learning has yet to be clearly explained, and realization of state-of-the-art performance with deep learning models requires tremendous amounts of labelled data. Further, optimization of deep learning models can require substantially long times for real-world applications. Hence, much effort is still needed to investigate deep learning theory and apply it in various challenging areas. This book looks …