Dendritic Computing: Branching Deeper into Machine Learning. (1st May 2022)

• Record Type: Journal Article
• Authors: Acharya, Jyotibdha; Basu, Arindam; Legenstein, Robert; Limbacher, Thomas; Poirazi, Panayiota; Wu, Xundong
• Is Part Of: Neuroscience. Volume 489(2022)
• Page Start: 275
• Page End: 289
• Publication Date: 2022-05-01
• Subjects: non-linear dendrites -- plasticity -- rewiring -- expressivity -- maxout networks -- machine learning -- deep neural networks; Neurochemistry -- Periodicals; Neurophysiology -- Periodicals; Neurology -- Periodicals; Neurochimie -- Périodiques; Neurophysiologie -- Périodiques; Neurochemistry; Neurophysiology; Electronic journals; Periodicals; Electronic journals; 612.8
• DOI: 10.1016/j.neuroscience.2021.10.001
• Languages: English
• ISSNs: 0306-4522

Abstract

Graphical abstract: Highlights: We review studies on computational implications of dendritic nonlinearities in neurons. We discuss how synaptic plasticity can interact with dendritic nonlinearities for improved learning. We link these studies to applications in machine learning and deep learning. Thus, we propose a path to more powerful and more efficient machine learning methods and hardware. Abstract: In this paper, we discuss the nonlinear computational power provided by dendrites in biological and artificial neurons. We start by briefly presenting biological evidence about the type of dendritic nonlinearities, respective plasticity rules and their effect on biological learning as assessed by computational models. Four major computational implications are identified as improved expressivity, more efficient use of resources, utilizing internal learning signals, and enabling continual learning. We then discuss examples of how dendritic computations have been used to solve real-world classification problems with performance reported on well known data sets used in machine learning. The works are categorized according to the three primary methods of plasticity used—structural plasticity, weight plasticity, or plasticity of synaptic delays. Finally, we show the recent trend of confluence between concepts of deep learning and dendritic computations and highlight some future research directions.