• Corpus ID: 7638149

A Technical Overview of the Neural Engineering Framework

@inproceedings{Stewart2012ATO,
  title={A Technical Overview of the Neural Engineering Framework},
  author={Terrence C. Stewart},
  year={2012}
}
The Neural Engineering Framework (NEF) is a general methodology that allows you to build largescale, biologically plausible, neural models of cognition [1]. In particular, it acts as a neural compiler: you specify the properties of the neurons, the values to be represented, and the functions to be computed, and it solves for the connection weights between components that will perform the desired functions. Importantly, this works not only for feed-forward computations, but recurrent connections… 

Figures from this paper

A spiking neural network of state transition probabilities in model-based reinforcement learning

This work provides an important first step toward understanding how a model-based system in the human brain could be implemented, and how this system contributes to human behaviour.

Think Fast: Time Control in Varying Paradigms of Spiking Neural Networks

This work constructs several network designs with varying degrees of biological plausibility and demonstrates their designs allow for a customized tradeoff between Biological plausibility, power efficiency, inference time, and accuracy.

BindsNET: A Machine Learning-Oriented Spiking Neural Networks Library in Python

It is argued that this package facilitates the use of spiking networks for large-scale machine learning problems and some simple examples by using BindsNET in practice are shown.

Implementation of the Neural Engineering Framework on the TrueNorth Neurosynaptic System

An implementation of the Neural Engineering Framework on IBM's TrueNorth Neurosynaptic system, where the crossbar array architecture itself, utilized in the TrueNorth hardware, can be used to compute the basic NEF calculations for any sized neural population, representing any dimensionality.

A spiking central pattern generator for the control of a simulated lamprey robot running on SpiNNaker and Loihi neuromorphic boards

This work proposes a spiking CPG neural network and its implementation on neuromorphic hardware as a means to control a simulated lamprey model and shows that this category of spiking algorithms shows a promising potential to exploit the theoretical advantages of neuromorphicHardware in terms of energy efficiency and computational speed.

Dynamics and Control in Spiking Neural Networks

Results show that classical control architectures, including, planning, feedback control, estimation, and learning, could indeed be enacted by a single recurrent spiking network with biophysically plausible dynamics, and provide a way to achieve distributed control over networks, which may confer robustness properties above and beyond traditional centralized control designs.

NeurOS™ and NeuroBlocks™ a neural/cognitive operating system and building blocks

Neuromorphic Models of the Amygdala with Applications to Spike Based Computing and Robotics

This thesis describes the creation of a spiking-neuron computational model of the amygdala, the brain region behind the authors' social interactions, and the simulation of the model using brain-inspired computer hardware, as well as the implementations of other spike-based computations on these hardwares.

Reproducibility and Rigour in Computational Neuroscience

The construction of a pipeline centered around an intermediate representation, termed dLEMS, which would facilitate simplified code generation not only for neuronal simulators but also for ODE-aware general purpose numerical platforms like Matlab or even C/Sundials, and the interplay between syntax and semantics is illustrated.

Reservoir Memory Machines as Neural Computers

This work achieves some of the computational capabilities of DNCs with a model that can be trained very efficiently, namely, an echo state network with an explicit memory without interference, which enables echo state networks to recognize all regular languages, including those that contractive echoState networks provably cannot recognize.

References

SHOWING 1-10 OF 16 REFERENCES

Neural Engineering: Computation, Representation, and Dynamics in Neurobiological Systems

The authors present three principles of neural engineering based on the representation of signals by neural ensembles, transformations of these representations through neuronal coupling weights, and the integration of control theory and neural dynamics, and argue that these guiding principles constitute a useful theory for generating large-scale models of neurobiological function.

Solving the Problem of Negative Synaptic Weights in Cortical Models

This work identifies a general form for the solution to the problem of converting unrealistic network models into biologically plausible models that respect this constraint and describes how the precise solution for a given cortical network can be determined empirically.

Spaun: A Perception-Cognition-Action Model Using Spiking Neurons

An overview of the Semantic Pointer Architecture: Unified Network (Spaun) model is presented and it is demonstrated that this biologically plausible spiking neuron model has the following features: Task Flexibility: No changes are made to the model between tasks.

A Unified Approach to Building and Controlling Spiking Attractor Networks

The addition of control shows how attractor Networks can be used as subsystems in larger neural systems, demonstrates how a much larger class of networks can be related to attractor networks, and makes it clear how attracting systems can be exploited for various information processing tasks in neurobiological systems.

Dynamic Behaviour of a Spiking Model of Action Selection in the Basal Ganglia Neural Structure

A detailed spiking neuron model of action selection is developed that takes into account a broad range of neurological details about the basal ganglia and is in accordance with the connectivity and neuron types found in this area.

A Neural Model of Rule Generation in Inductive Reasoning

A biologically plausible method is presented for accomplishing this task and implemented in a spiking neuron model that is able to generate the rules necessary to correctly solve Raven's items, as well as recreate many of the experimental effects observed in human subjects.

A Brain-Machine Interface Operating with a Real-Time Spiking Neural Network Control Algorithm

A Kalman-filter based decoder via a spiking neural network (SNN) and tested it in brain-machine interface (BMI) experiments with a rhesus monkey, holding promise for hardware SNN implementations of statistical signal processing algorithms on neuromorphic chips.

The Attentional Routing Circuit: Receptive Field Modulation Through Nonlinear Dendritic Interactions

The Attentional Routing Circuit is presented, which provides an explanation of RF modulation as well as testable predictions about nonlinear cortical dendrites and attentional changes of receptive field properties.

Learning to Select Actions with Spiking Neurons in the Basal Ganglia

This model is compared to animal data in the bandit task, which is used to test rodent learning in conditions involving forced choice under rewards and indicates a good match in terms of both behavioral learning results and spike patterns in the ventral striatum.

Vector Symbolic Architectures answer Jackendoff's challenges for cognitive neuroscience

This paper claims that a little-known family of connectionist models (Vector Symbolic Architectures) are able to meet Jackendoff's challenges and is able to answer the challenges answered by some technical innovation in connectionist modelling.