Neuromorphic computing seeks to replicate aspects of the structure and function of biological neural networks in electronics. The aim is to develop machines that will ultimately display the same capabilities as the human brain, including its incredible energy efficiency.¹

Future Horizons:

×××

5-yearhorizon

Useful neuronal architectures emerge

New memory-based hardware better suited to implementing spiking neural networks reaches commercial scale thanks largely to demand for more energy-efficient chips for deep learning. Algorithms for navigation, trajectory generation and motor control create building blocks that can be deployed in small robots, but only in research settings as they still underperform deep learning. Neuromorphic technologies that mimic how the brain processes vision and smell show promise.

10-yearhorizon

Neuromorphic computing is developed for embodied AI

More sophisticated models of neurons and a deeper understanding of plasticity rules boost the capabilities of neuromorphic architectures. Deep-learning AI models are increasingly adapted to work on neuromorphic hardware to take advantage of energy savings. Neuro-emulators carry out brain-scale simulations of bio-inspired architectures, advancing fundamental understanding of brain dynamics. Neuromorphic engineers understand how to combine algorithmic building blocks to accomplish complex tasks, and the approach becomes the dominant computing framework for embodied AI that works with sensory signals and motion control. Low-power neuromorphic edge devices are used to pre-process all kinds of sensor data.

25-yearhorizon

Experiments demonstrate brain-like memory and logic

We have a better understanding of the processes and causal relationships between different levels of the brain, going from molecular to cellular to circuit to structures. We map the connectivity and structure, as well as the genetic programs and rules of plasticity by which the structure is formed. Various experimental realisations of neuromorphic computing demonstrate memory and logic that, while still primitive compared to the naturally evolved brain, work in recognisably mammalian ways. Robots powered by neuromorphic computing are widely deployed in the defence and care sectors. AI is pervasive and an integral part of our environment, with the technological basis of AI shifted to neuromorphic technologies.

This requires the development of new kinds of computer chips that more faithfully mimic the way neurons work and are arranged.² ³ One particular challenge is to replicate how brain cells communicate via spikes of electrical activity, which is very different from how conventional processors operate. Current models of biological neurons and learning rules are also too simplistic, brushing over physiological details known to play an important role in the brain.

Nonetheless, large-scale neuromorphic systems are coming online. Intel’s new Hala Point system combines 1,152 of its Loihi 2 neuromorphic processors to emulate up to 1.15 billion neurons.⁴ And SpiNNcloud Systems is selling neuromorphic supercomputers based on chip designs pioneered by the SpiNNAker project at the University of Manchester, UK.⁵ ⁶

There have also been breakthroughs in chips that carry out computation directly in memory.⁷ These are well-suited to implementing neuromorphic computing, but also holds promise for reducing deep-learning AI’s energy consumption. That could lead to growing convergence between the two fields.

However, biological realism is still limited. Breakthroughs in mapping neural connections in real brains and more biologically plausible learning approaches are helping close the gap.⁸ ⁹ But faithfully replicating the brain’s function and structure in silicon remains a distant goal.

Neuromorphic Computing - Anticipation Scores

The Anticipation Potential of a research field is determined by the capacity for impactful action in the present, considering possible future transformative breakthroughs in a field over a 25-year outlook. A field with a high Anticipation Potential, therefore, combines the potential range of future transformative possibilities engendered by a research area with a wide field of opportunities for action in the present. We asked researchers in the field to anticipate:

The uncertainty related to future science breakthroughs in the field
The transformative effect anticipated breakthroughs may have on research and society
The scope for action in the present in relation to anticipated breakthroughs.

This chart represents a summary of their responses to each of these elements, which when combined, provide the Anticipation Potential for the topic. See methodology for more information.