Continual Learning of Multiple Cognitive Functions with Brain-inspired Temporal Development Mechanism

Continual Learning of Multiple Cognitive Functions with Brain-inspired Temporal Development Mechanism Bing Han 1, 3,#, Feifei Zhao 1, #, Yinqian Sun 1, Wenxuan Pan 1,3, Yi Zeng 1, 2,3,4, 1 Brain-inspired Cognitive Intelligence Lab, Institute of Automation,Chinese Academy of Sciences 2 State Key Laboratory of Brain Cognition and Brain-inspired Intelligence Technology, Chinese Academy of Sciences 3 School of Artificial Intelligence, University of Chinese Academy of Sciences 4 Center for Long-term Artificial Intelligence Corresponding authors: yi.zeng@ia.ac.cn # Co-first authors with equal contribution April 9, 2025 Abstract Cognitive functions in current artificial intelligence networks are tied to the exponential increase in network scale, whereas the human brain can continuously learn hundreds of cognitive functions with remarkably low energy consumption. This advantage is in part due to the brain's cross-regional temporal development mechanisms, where the progressive formation, reorganization, and pruning of connections from basic to advanced regions, facilitate knowledge transfer and prevent network redundancy. Inspired by these, we propose the Continual Learning of Multiple Cognitive Functions with Brain-inspired Temporal Development Mechanism(TD-MCL), enabling cognitive enhancement from simple to complex in Perception-Motor-Interaction(PMI) multiple cognitive task scenarios. The TD-MCL model proposes the sequential evolution of long-range connections between different cognitive modules to promote positive knowledge transfer, while using feedback-guided local connection inhibition and pruning to effectively eliminate redundancies in previous tasks, reducing energy consumption while preserving acquired knowledge. Experiments show that the proposed method can achieve continual learning capabilities while reducing network scale, without introducing regularization, replay, or freezing strategies, and achieving superior accuracy on new tasks compared to direct learning. The proposed method shows that the brain's developmen-1 arXiv:2504.05621v1 Keywords Brain-inspired Temporal Development, Multiple Cognitive Functions Continual Learning, Evolutionary Growth Long-range Connectivity, Feedback-guided Suppression and Pruning, Biological Synaptic Plasticity 1 Introduction Artificial intelligence algorithms have achieved remarkable success across various fields, but their enhancement of cognitive functions relies on the massive stacking of parameters, often facing challenges in balancing memory capacity with energy consumption[1]. In contrast, the brain requires only 20 watts of power to gradually master a rich array of cognitive functions during its developmental process, offering valuable biological insights.