Cognitive Twin Native Systems: From Digital Representation to Task-Oriented Intelligence

Artificial intelligence, digital twins, communication networks, and distributed systems have significantly advanced modern intelligent services. However, most existing systems remain optimized around intermediate objectives, such as prediction accuracy, synchronization fidelity, and communication efficiency, rather than the ultimate goal of task accomplishment. This DL introduces Cognitive Twin Native Systems (CTNS), a new paradigm that places task-oriented intelligence at the center of system design. Moving beyond conventional Digital Twins, CTNS tightly integrates representation, learning, cognition, decision-making, and task execution within a continuously evolving closed-loop architecture. The talk discusses the evolution from Digital Twins to Human Digital Twins and Cognitive Digital Twins, and presents five foundational capabilities required for building Cognitive Twin systems: predictive digital foundations, sustainable cognitive learning, collaborative intelligence, distributed intelligence, and system-level cognitive optimization. Finally, the talk explores future opportunities at the intersection of networking, artificial intelligence, optimization, and digital twin technologies, and discusses how AI-native Cognitive Digital Twins may enable the next generation of intelligent systems. The central message is simple: the future is not merely about representing the world more accurately, but about building systems that can understand, reason, decide, and act in pursuit of human objectives.