Abstract: Collaborative networks have become an important resource channel for firms involved in exploratory innovation. While research on the unidirectional relationship between the two has matured, there is limited discussion on their co-evolution and the factors influencing this dynamic. Based on a network capital perspective, this paper constructs a multilayer network using the invention patent data of 145 firms in the field of artificial intelligence, from 2002 to 2022. It examines the co-evolutionary mechanism between exploratory innovation and collaborative network through the stochastic actor-oriented model, exploring the effects of the inventor network, knowledge network, and the ego-network of firms. The findings show a co-evolutionary relationship between exploratory innovation and collaborative networks, shaped by social selection and social influence. In this co-evolution context, the positive impact of the small-world structure in the inventor network on exploratory innovation, as well as its negative impact on collaborative network, is mitigated through neutralization. Additionally, this paper highlights that the positive effects of knowledge network size, firm ego-network efficiency, and recombinant distance on both exploratory innovation and collaborative network expansion are further strengthened. These results provide a novel approach and perspective for research on innovation networks and have practical implications for firms engaged in collaborative innovation activities.

Key words: exploratory innovation, collaborative network, co-evolution, network capital, multilayer network

摘要： 合作网络已成为企业从事探索式技术创新的重要资源渠道，且两者之间的单向关系研究趋于成熟，但尚缺乏对两者协同演化及其动因的讨论。基于网络资本视角，利用145家人工智能企业在2002~2022年间的发明专利数据构建多层网络，借助随机行动者导向模型验证了探索式创新与合作网络的协同演化机制，并探究了发明人网络、知识网络和企业自我网络特征的影响。研究表明：在社会选择机制与社会影响机制作用下，探索式创新与合作网络存在相互促进的协同演化关系；在协同演化情境中，发明人网络小世界性对探索式创新的积极影响和对合作网络扩张的消极影响会发生“中和”而进一步减弱，知识网络规模、企业自我网络效率以及重组距离对探索式创新与合作网络扩张的正向作用则进一步增强。研究结果为创新网络研究提供了新的方法和视角，对企业从事协同创新活动具有现实意义。

关键词: 探索式创新, 合作网络, 协同演化, 网络资本, 多层网络