Search papers, labs, and topics across Lattice.
This paper introduces ComAI, a paradigm for converging communication and AI in next-generation wireless networks, drawing inspiration from natural intelligence processes. The authors propose a ComAI framework encompassing information sensing, intellicise semantic communication networks for information cognition, and large-scale models for information decision-making. Experimental validations from a 6G test network and potential applications like embodied AI robots are presented to support the feasibility and impact of ComAI.
ComAI promises to revolutionize wireless networks by merging communication and AI, enabling intelligent and concise semantic communication that simplifies signal processing and network organization.
The convergence of communication and artificial intelligence (AI) is transforming next-generation wireless networks. Inspired by natural intelligence, particularly in how data, information, and knowledge are processed for cognition, reasoning, learning, decision-making, and environmental interaction, we introduce a new paradigm called ComAI. This paradigm is expected to play a key role in future wireless networks. In this paper, we highlight the motivations behind the convergence of communications and AI and present a roadmap that outlines the evolution of wireless communications alongside advancements in AI. Then, we introduce the key idea of ComAI and provide a general framework of ComAI, which includes information sensing and data collection, information cognition through intellicise semantic communication networks, and information decision with large-scale models. In particular, intellicise semantic communication networks refer to intelligent and concise communication systems that leverage semantic-level information to simplify and restructure signal processing, information cognition, and network organization. We further discuss the foundational theories and key techniques for the convergence of communication and AI. Additionally, we provide experimental validations, including insights from the 6G test network, and highlight potential application scenarios such as embodied AI robots and multi-agent collaborative communications. Finally, we outline the future directions and open issues in the development of ComAI, sharing our thoughts and insights on potential solutions.
