Abstract:Spiking neural networks were introduced to understand spatiotemporal information processing in neurons and have found their application in pattern encoding, data discrimination, and classification. Bioinspired network architectures are considered for event-driven tasks, and scientists have looked at different theories based on the architecture and functioning. Motor tasks, for example, have networks inspired by cerebellar architecture where the granular layer recodes sparse representations of the mossy fiber (MF) inputs and has more roles in motor learning. Using abstractions from cerebellar connections and learning rules of deep learning network (DLN), patterns were discriminated within datasets, and the same algorithm was used for trajectory optimization. In the current work, a cerebellum-inspired spiking neural network with dynamics of cerebellar neurons and learning mechanisms attributed to the granular layer, Purkinje cell (PC) layer, and cerebellar nuclei interconnected by excitatory and inhibitory synapses was implemented. The model's pattern discrimination capability was tested for two tasks on standard machine learning (ML) datasets and on following a trajectory of a low-cost sensor-free robotic articulator. Tuned for supervised learning, the pattern classification capability of the cerebellum-inspired network algorithm has produced more generalized models than data-specific precision models on smaller training datasets. The model showed an accuracy of 72%, which was comparable to standard ML algorithms, such as MLP (78%), Dl4jMlpClassifier (64%), RBFNetwork (71.4%), and libSVM-linear (85.7%). The cerebellar model increased the network's capability and decreased storage, augmenting faster computations. Additionally, the network model could also implicitly reconstruct the trajectory of a 6-degree of freedom (DOF) robotic arm with a low error rate by reconstructing the kinematic parameters. The variability between the actual and predicted trajectory points was noted to be ± 3 cm (while moving to a position in a cuboid space of 25 × 30 × 40 cm). Although a few known learning rules were implemented among known types of plasticity in the cerebellum, the network model showed a generalized processing capability for a range of signals, modulating the data through the interconnected neural populations. In addition to potential use on sensor-free or feed-forward based controllers for robotic arms and as a generalized pattern classification algorithm, this model adds implications to motor learning theory.

A Sparsely Distributed and Statistically Independent Representation of Parallel Fiber Coding in Cerebellum

C^2:Co-design of Robots via Concurrent Networks Coupling Online and Offline Reinforcement Learning

A Simplified Cerebellar Model with Priority-based Delayed Eligibility Trace Learning for Motor Control

A Simplified Cerebellum-Based Model for Motor Control in Brain Based Devices.

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

Real-time Implementation of the Cerebellum Neural Network

A cerebellar population coding model for sensorimotor learning

Learning Sparse Spatial Codes for Cognitive Mapping Inspired by Entorhinal-Hippocampal Neurocircuit

Sparse Coding in a Dual Memory System for Lifelong Learning

Population coding in the cerebellum: a machine learning perspective

Cortico-cerebellar networks as decoupling neural interfaces

A cerebellum inspired spiking neural network as a multi-model for pattern classification and robotic trajectory prediction

Cooperative coding of continuous variables in networks with sparsity constraint

Competitive learning to generate sparse representations for associative memory

CerebelluMorphic: Large-Scale Neuromorphic Model and Architecture for Supervised Motor Learning

Online Discriminative Sparse Coding As a Systems-level Model of the Primate's Dorsal and Ventral Pathways

The Relative Advantages of Sparse Versus Distributed Encoding for Associative Neuronal Networks in the Brain

Adaptive Robotic Control Driven by a Versatile Spiking Cerebellar Network

Relative representations for cognitive graphs

Organizing Principles of Real-Time Memory Encoding: Neural Clique Assemblies and Universal Neural Codes

Cerebellar Functional Parcellation Using Sparse Dictionary Learning Clustering.