Neuromorphic computing in spiking neuron-astrocyte networks
summary
Models biological astrocyte calcium as third-order nonlinear dynamics coupled by diffusion to spiking neurons (Hodgkin‑Huxley style astrocyte equations, Izhikevich neurons), shows seconds-scale calcium signals that integrate local synaptic activity and release gliotransmitters to modulate weights, implements spike-timing-dependent plasticity plus astrocyte modulation, stores overlapping patterns with graded amplitude encoding, and reports a 15% classification uplift on MNIST-like tests; builds multi-compartment and network simulations and suggests memristive neuroelectronics as hardware path.
AI-written summary · about
chapters (10)
- 0:00Neuromorphic computing intro
- 1:07Biologically relevant neural modeling
- 4:42Biological vs artificial networks
- 8:23Calcium dynamics in astrocytes
- 13:48Astrocyte-neuron interaction
- 15:30Astrocytes in working memory
- 18:04Biologically relevant memory storage
- 20:54Performance gains with astrocytes
- 21:38Team and conclusion
- 22:01Question and answer session
related talks
01h 13mComputation in Networks of Neurons in the Brain II
51mThe Brain as an Efficient, Robust Adaptive Learner
01h 57mIntroduction to Neuromorphic Computing
41mBrain-inspired computational primitives for reliable neuromorphic intelligence
43mKeeping Moore's law alive: Neuromorphic computing
36mBrain's Neural Network as a service
47mDynamic Neural Network Structures Through Stochastic Rewiring
30mCracking the code: neuronal networks in the brain