brain inspired models Flashcards
(30 cards)
What is the goal of brain-inspired models in AI?
To replicate cognitive functions and learning mechanisms of the human brain.
What makes the human brain inspiring for AI?
Its ability to generalize, learn from few examples, and operate efficiently.
What are the key components of a biological neuron?
Dendrites, soma, axon, and synapses.
What triggers an action potential in a neuron?
Depolarization past a threshold due to ion flow.
What happens at a synapse during signal transmission?
Neurotransmitters cross the synaptic cleft and bind to receptors.
How does an artificial neuron differ from a biological neuron?
It outputs continuous values and lacks time dynamics.
What kind of signals do spiking neural networks use?
Discrete spike events over time.
What was the McCulloch-Pitts neuron model?
A binary model that fired based on thresholded weighted inputs.
What is the Hodgkin-Huxley model known for?
Modeling action potential using electrical circuit analogies.
What does the HH model treat the membrane as?
A capacitor with ion channels as variable resistors.
What does the membrane voltage equation in HH include?
Capacitive current and ionic currents from K+, Na+, and leak channels.
What do the variables m, h, and n represent in HH models?
Gating variables that control ion channel opening and closing.
What is the main drawback of the HH model?
It is computationally expensive for large-scale networks.
What does the Izhikevich model aim to balance?
Biological realism and computational efficiency.
What can the Izhikevich model reproduce?
Diverse neuron firing patterns with low computational cost.
What is the Blue Brain Project?
A project to simulate a neocortical column at cellular resolution.
What brain regions are modeled in the PFC–DRN interaction model?
Prefrontal cortex and dorsal raphe nucleus.
What is neuromorphic computing?
Hardware designed to mimic neural computation and communication.
What platform uses ARM processors for spiking simulation?
Spinnaker (University of Manchester).
What is BrainScaleS designed for?
Analog modeling of synaptic dynamics and neural circuits.
What are benefits of neuromorphic platforms?
Low power, real-time simulation, and massive parallelism.
What is PyNN used for?
A Python interface for multiple neural simulators.
What simulator is best for detailed neuron modeling?
NEURON (uses HOC for scripting).
What simulator handles large simplified neuron networks?
NEST.
