Abstract:Brain-inspired computing uses the law of brain processing information to simulate the brain. The network formed by the transmission of different signals between neurons has the ability to deal with complex nonlinear problems. This is an important means of artificial intelligence implementation. As the model basis of brain-inspired computing research, spiking neural network has powerful perception, learning and decision ability. The multi-compartment model of basal dendrites and apical dendrites processes and integrates the input information at different levels of the cell body, which is more similar to the behavior of biological neurons. Context-dependent learning is a process in which research subjects respond differently to different environmental conditions and use situational information to make behavioral changes. The higher cognitive function of human brain is simulated according to the task model. Our research shows that the properties of neuronal models with dendrites and soma compartments play an important role in the performance of context-dependent learning. The results show that the dendrite conductance in the dendritic compartments affect the regeneration process of dendrite potentials, affect the integration of dendrites into neuronal inputs, and thus change the learning performance of the network. The reversal potential in dendrite compartments and the coupling coefficient between dendrites and cell bodies guide the decision-making process of cell input current, affect the change of neuronal activity, and influence the accuracy of network learning. Moreover, we compare the effects of apical dendrite and basal dendrite compartments on network learning performance, and reveal the differences between different types of compartments on network learning ability. This work can be used for smart security, unmanned environmental awareness systems and intelligent production lines.

Effect Exploration of the Spiking Neuron Property and Heterogeneity on Brain-Inspired Context-Dependent Learning

Investigation on the Effect of Spiking Neuron Properties with Dendrite Compartments in Brain-inspired Context-dependent Learning

Toward Efficient Processing and Learning with Spikes: New Approaches for Multispike Learning

Enhancing learning in spiking neural networks through neuronal heterogeneity and neuromodulatory signaling

Unleashing the Potential of Spiking Neural Networks for Sequential Modeling with Contextual Embedding.

Context Gating in Spiking Neural Networks: Achieving Lifelong Learning through Integration of Local and Global Plasticity

Exploiting Noise As a Resource for Computation and Learning in Spiking Neural Networks

Deep CovDenseSNN: A Hierarchical Event-Driven Dynamic Framework with Spiking Neurons in Noisy Environment

Coordination and Compensation of Neuronal Firing Sensitivity and Leakage Rate for Brain-Inspired One-Shot Learning

A Brain-Inspired Spiking Neural Network Model with Temporal Encoding and Learning

Dive into the Power of Neuronal Heterogeneity

Effects Investigation of Critical Spiking Neuronal Feature on Dendritic Credit Assignment for Efficient Spike-based Learning

Biologically Inspired Heterogeneous Learning for Accurate, Efficient and Low-Latency Neural Network

Learning Improvement of Spiking Neural Networks with Dynamic Adaptive Hyperparameter Neurons

Toward stochastic neural computing

Beyond spiking networks: The computational advantages of dendritic amplification and input segregation

Advanced Computing and Related Applications Leveraging Brain-inspired Spiking Neural Networks

Heterogeneous Neuronal and Synaptic Dynamics for Spike-Efficient Unsupervised Learning: Theory and Design Principles

Brain-inspired Chaotic Spiking Backpropagation

How Neuronal Noises Influence the Spiking Neural Networks’s Cognitive Learning Process: A Preliminary Study

Memory-Dependent Computation and Learning in Spiking Neural Networks Through Hebbian Plasticity