Computer Neurons

In this week’s Science magazine, D. Modha of IBM describes a computer chip based on the organization of the mammal brain, with a million computational units, or neurons. Each of these connects to 256 others. His work recalls Spike Gerrell’s brilliant cartoons from last year’s Economist magazine.

20130803_STD001_0      20130803_STD002_0

Modha’s chip is a heavily interconnected structure of the biologically-inspired, “neuromorphic” systems; these are an efficient way of handling a lot of data at the same time integrating computation, communication and memory.

Instead of binary ones and zeros, the units of computation here are spikes. A neuron generates a spike and sends it across the chip to other neurons. Software has to be written completely differently for such spiking-network systems.

939.jpg   _76801184_ibm_synapse_20140807_024  _76801177_ibm_synapse_20140807_007

Dr Modha and his team managed to engineer an interconnected 64-by-64 grid of these cores on to a single chip, delivering over one million neurons in total. Because each neuron is connected to 256 others, there are more than 256 million connections or “synapses”.

Biological neurons, packed inside the brain, send and receive something in the order of 10,000 connections each.

_76801179_untitled                            20130803_STD003_0

Each of 4,096 “neurosynaptic cores” on the chip contains 256 neurons, which connect to 256 other neurons within and outside that core; this diagram illustrates links between just 64 cores.

(cartoons by Spike Gerrell in The Economist, August 3 2013)