Technology Fitness Landscape for Design Innovation: A Deep Neural Embedding Approach Based on Patent Data

In the past decade, artificial intelligence, cloud computing, quantum computing, and 5G communication technologies undergo rapid advances. Meanwhile, tremendous innovations also emerge and gain momentum in traditional technological domains, such as autonomous vehicles [1], drug discovery [2], and protein structure prediction [3]. Such contemporary innovation phenomena call for new theories and frameworks to explain them, understand the driving forces, and inform future innovation. Many contemporary design innovations share one characteristic in common: they are based on the synthesis and fusion of different technological domains, which used to be unrelated and separately developed, e.g., artificial intelligence and automobile. The rise of such innovations has ambiguated the boundaries of technological domains and industries.