Regional knowledge network is an important way to explain how innovation takes place in regions. There has been a large body of literature addressing the properties of scientific or technological collaboration networks in different regions. However, most of the studies treat science and technology separately and focus less on their interaction. A well-established stream of literature has demonstrated that science is a critical source of knowledge in the development of technology. Promoting the knowledge flow from science to technology helps gain competitive advantage in emerging industries. However, little evidence shows how science is transferred to technology through actors in regions. Therefore, this paper aims to identify the role of different local actors in regional knowledge network from the perspective of science-technology linkage. In this way, policy makers can optimize the layouts of entities and develop science and technology hubs in city clusters so that science and technology can be more deeply interacted, which contributes to high-quality regional development. Patent-to-paper citations are commonly interpretated as the knowledge diffusion from science to technology.

Therefore, we collect patent-to-paper citations from 2001 to 2020, which involves the organizations located in the Beijing-Tianjin-Hebei, Yangtze River Delta, and Pearl River Delta regions respectively to establish the inter-organizational knowledge networks in the regions. The characteristics of local actors in the knowledge networks of three city clusters are compared at the level of the entire network and ego - network with the help of social network and correlation analysis. The results are: (1) Universities play a significant role as the knowledge producer, absorber and brokers. Research Institutions play a similar but less important role. Large and medium-sized enterprises and small and micro-enterprises focus more on knowledge absorption, with only a small portion as both the knowledge producer and broker. The ability of small and micro - enterprises in knowledge production, absorption and brokerage is weakest among all types of actors. (2) In the knowledge networks of the Beijing-Tianjin-Hebei region and the Yangtze River Delta, universities’ ability in knowledge output, is in line with their capability in knowledge production, and brokerage. For large and medium-sized enterprises and small and micro-enterprises, the ability shows little correlation. In the Pearl River Delta knowledge network, universities mostly focus on either knowledge absorption or output, while the ability of enterprises in knowledge output, input, and brokerage is highly correlated. (3) In the Beijing-Tianjin-Hebei region and the Yangtze River Delta networks, universities have the closest communication with external entities, while in the Pearl River Delta network, large and medium - sized enterprises have the highest level of direct communication with external entities. This paper contributes to the theoretical literature in two ways. First, we add the understanding of the role of different actors in regional knowledge networks. Specifically, our results reveal the heterogeneous characteristics of actors in their involvement in transferring knowledge from science to technology. Second, we provide a novel approach of assessment for the actors in regional innovation system based on social network analysis with patent-to-paper citations. Our findings have policy implications for improving regional innovation systems with various actors.