The paper "A Collective AI via Lifelong Learning and Sharing at the Edge" explores the concept of creating a collective artificial intelligence (AI) system where multiple AI units can learn independently over their lifetimes and share their knowledge with each other. This vision aims to create a society of AI systems that can contribute to and benefit from collective knowledge, enhancing their performance and adaptability. The authors highlight the importance of incremental learning, knowledge exchange, and the use of both local data and communication for learning. They discuss the challenges and opportunities in developing such a system, including the need for scalable and resilient hardware platforms, effective knowledge sharing protocols, and the optimization of computational resources. The paper also reviews recent advances in lifelong learning, federated learning, and distributed systems, and outlines potential application areas such as multi-agent active sensing, space exploration, personalized medicine, and distributed cyber-security systems. The authors conclude by identifying key challenges, such as scalability, communication protocols, and computational efficiency, that need to be addressed to realize the full potential of a collective AI system.The paper "A Collective AI via Lifelong Learning and Sharing at the Edge" explores the concept of creating a collective artificial intelligence (AI) system where multiple AI units can learn independently over their lifetimes and share their knowledge with each other. This vision aims to create a society of AI systems that can contribute to and benefit from collective knowledge, enhancing their performance and adaptability. The authors highlight the importance of incremental learning, knowledge exchange, and the use of both local data and communication for learning. They discuss the challenges and opportunities in developing such a system, including the need for scalable and resilient hardware platforms, effective knowledge sharing protocols, and the optimization of computational resources. The paper also reviews recent advances in lifelong learning, federated learning, and distributed systems, and outlines potential application areas such as multi-agent active sensing, space exploration, personalized medicine, and distributed cyber-security systems. The authors conclude by identifying key challenges, such as scalability, communication protocols, and computational efficiency, that need to be addressed to realize the full potential of a collective AI system.