Dendritic learning is inspired by the brain’s complex dendritic structures and biologically plausible learning mechanisms, and it exhibits substantial potential as a next-generation framework for biologically plausible artificial intelligence. Over the years, sustained research efforts have positioned dendritic learning as a promising and rapidly advancing direction in AI research. This paper presents a comprehensive survey of dendritic learning, encompassing its architectures, learning algorithms, and application domains. The main contents of this survey include: the neuroscience fundamentals of dendrites, a systematic review of existing dendritic learning architectures from the perspectives of dendritic plasticity and morphology, biologically plausible dendritic learning rules, a summary of real-world applications of dendritic learning, and an exploration of open questions and potential research directions in this field.
Gu et al. (Wed,) studied this question.