neuromorphic computing
In plain wordsComputer chips designed to work more like brains, with artificial neurons that send brief spikes only when needed, aiming to do AI tasks with far less energy.
Computing hardware and systems modeled on the structure and dynamics of biological nervous systems: many simple neuron-like units that communicate by sparse, asynchronous spikes, with memory (synaptic weights) co-located with computation and activity driven by events rather than a global clock. Implementations range from analog and mixed-signal circuits to large digital many-core chips, typically running spiking neural networksspiking neural network, n.: A neural network that, like the brain, communicates with brief pulses, or spikes, sent only when a neuron's charge crosses a threshold. On special chips this can save a great deal of energy.; examples include IBM's TrueNorth (2014; 1 million neurons and 256 million synapses), Intel's Loihi (2017) and Loihi 2 (2021), the University of Manchester's SpiNNaker and Heidelberg's BrainScaleS. Proponents cite large energy savings for sparse, low-latency workloads such as sensing and robotics; critics note that spiking models still trail conventional deep learningdeep learning, n.: Machine learning done with neural networks built from many stacked layers: raw data such as pixels, sound or text goes in at one end, and each layer turns what it receives into something more abstract. in accuracy and tooling, so neuromorphic chips remain largely research platforms.
Origin
Carver Mead of Caltech developed the concept in the late 1980s, building analog VLSI circuits that mimicked neural and retinal processing, and is credited with the term; his paper 'Neuromorphic electronic systems' (Proceedings of the IEEE, October 1990) set out the program and is its standard early citation.
Related
- spiking neural networkspiking neural network, n.: A neural network that, like the brain, communicates with brief pulses, or spikes, sent only when a neuron's charge crosses a threshold. On special chips this can save a great deal of energy.
- analog in-memory computinganalog in-memory computing, n.: Doing AI math inside the memory itself, using electrical currents through stored values, instead of shuttling numbers back and forth to a separate processor. It could save lots of energy but is imprecise.
- von Neumann bottleneckvon Neumann bottleneck, n.: The traffic jam caused by ordinary computers keeping data and processing in separate places, so information must constantly shuttle back and forth through one narrow channel.
- AI acceleratorAI accelerator, n.: Any chip built or tuned to speed up AI — graphics chips, Google's TPUs, phone 'neural engines' and many startup designs. Such chips do the heavy math of training and running models.
- hardware lotteryhardware lottery, n.: The idea that research ideas succeed partly by luck — because they happen to suit the computer chips available at the time — rather than purely because they are better.
- GPUGPU, abbr.: A computer chip first designed to draw video-game graphics that turned out to be ideal for AI, because it can do thousands of simple calculations at the same time.
Sources
- Mead, C. (1990). Neuromorphic electronic systems. Proceedings of the IEEE 78(10), 1629–1636.
- Merolla, P. A. et al. (2014). A million spiking-neuron integrated circuit with a scalable communication network and interface. Science 345(6197).
Mentioned in
AI acceleratorAI accelerator, n.: Any chip built or tuned to speed up AI — graphics chips, Google's TPUs, phone 'neural engines' and many startup designs. Such chips do the heavy math of training and running models., Hebbian learningHebbian learning, n.: The idea that a connection between two brain cells grows stronger when they are active at the same time — the principle behind the slogan 'cells that fire together wire together'., analog in-memory computinganalog in-memory computing, n.: Doing AI math inside the memory itself, using electrical currents through stored values, instead of shuttling numbers back and forth to a separate processor. It could save lots of energy but is imprecise., cognitive computingcognitive computing, n.: A label, promoted heavily by IBM in the 2010s, for computer systems that learn, reason and interact with people in humanlike ways. It was largely a marketing term and has given way to 'AI'., connectomeconnectome, n.: A full chart of how a nervous system is wired: which nerve cells, or which brain regions, are linked to which others., hardware lotteryhardware lottery, n.: The idea that research ideas succeed partly by luck — because they happen to suit the computer chips available at the time — rather than purely because they are better., photonic computingphotonic computing, n.: Using light instead of electricity to compute or move data. Light can carry huge amounts of information with little energy, but building full computers from it is still very hard., quantum machine learningquantum machine learning, n.: Research into whether quantum computers could learn from data faster or better than ordinary computers, and into using AI to study quantum systems. Practical advantages have not yet been shown., reservoir computingreservoir computing, n.: A shortcut way to build a network for time-varying signals: a large, fixed, randomly wired network echoes the input, and only a simple final readout is trained., spiking neural networkspiking neural network, n.: A neural network that, like the brain, communicates with brief pulses, or spikes, sent only when a neuron's charge crosses a threshold. On special chips this can save a great deal of energy., von Neumann bottleneckvon Neumann bottleneck, n.: The traffic jam caused by ordinary computers keeping data and processing in separate places, so information must constantly shuttle back and forth through one narrow channel., whole brain emulationwhole brain emulation, n.: Copying a particular person's brain into a computer in such detail, down to individual neurons and their connections, that the simulation would think and behave like that person.