Integrating Generative AI with Network Digital Twin for 6G: An Edge-Cloud Collaborative Approach

The sixth-generation (6G) of wireless systems aim to support the emerging Internet of Everything (IoE) applications and provide Artificial Intelligence-Generated Content (AIGC) services to diverse mobile devices. The strict communication requirements of various IoE applications and the personalized user requirements of mobile AIGC services pose great challenges to the 6G network management. Creating a digital twin (DT) to represent the wireless networks can facilitate the achievement of zero-touch and autonomous wireless network management with accurate modeling of the physical network and proactive analysis of its behavior. In this article, a hybrid deployment architecture of a joint DT for 6G network is introduced, integrating with the generative artificial intelligence (GenAI) models which are trained and deployed in an edge-cloud collaborative way. A integration framework between GenAI and network DT is presented to realize a cycle of benefits. GenAI models offer efficient solutions for ubiquitous intelligent decision-making during the life cycle of DT while the joint DT provides assistance in mobile AIGC services provisioning to reduce the resource consumption and improve user experience. A case study is presented for illustrating how to utilize DT capabilities to enhance the computing resource allocation for fine-tuning and inference, providing personalized AIGC services.