In part, the interest in unconventional computing has been driven by a growing appreciation that information-processing is a fundamental part of many natural processes and that clever engineering can help harness these processes for our own use. Efforts to exploit the unique behaviour of quantum systems — covered elsewhere in the Radar — have achieved the most attention and investment. But a host of other unconventional approaches to computing show promise.
The unique capabilities of the human brain have drawn considerable attention, for example. Today’s AI software is already loosely inspired by the way neurons work, but neuromorphic engineers are attempting to go a step further and create hardware that more faithfully replicates the brain’s form and function. Others are investigating whether small conglomerations of brain cells known as organoids can be coaxed into solving computational tasks.
Even simpler biological processes may hold enormous computing potential. DNA’s ability to encode, transform and replicate information holds clear parallels to the binary code at the heart of classical computers, and raises the possibility of controlling more complex cellular and intercellular processes that could enable us to go beyond conventional circuit-based computing paradigms.
The prospect of encoding computations in light rather than electricity is also showing significant promise. The approach could offer substantial speed-ups on a variety of problems, including AI, and commercialisation is already well under way.
KEY TAKEAWAYS
Although the standard pathways of computing technology have created a revolution in the human experience, there are other ways of processing information that might offer radically different possibilities. One is Neuromorphic computing, where technology aims to mimic the architectures and processes, where they are understood, of the biological brain. Given evolution’s role in developing biological brains, this has the potential to be fast and energy-efficient. Also in progress is an exploration of Organoid intelligence, where biological brain tissue is grown at very small scale and its signal and information-processing capabilities explored by in vitro experiments. In Cellular computing, the biological machinery inside networks of cells could one day be harnessed to solve challenges in areas as varied as environmental remediation and drug discovery. This involves computing with whole molecules as the input and output, and has the potential to perform a variety of tasks that are beyond the capabilities of conventional computing. On a different tack, researchers are experimenting with replacing the electrons of traditional computing with the photons of Optical computing. This has several potential advantages, including high data throughput and improved energy efficiency.
Three fundamental questions guide GESDA’s mission and drive its work: Who are we, as humans? How can we all live together? How can we ensure the well-being of humankind and the sustainable future of our planet? We asked researchers from the field to anticipate what impact future breakthroughs could have on each of these dimensions. This wheel summarises their opinions when considering each of these questions, with a higher score indicating high anticipated impact, and vice versa.
This requires the development of new kinds of computer chips that more faithfully mimic the way neurons work and are arranged.2,3 In particular, neuromorphic systems replicate how brain cells communicate via spikes of electrical activity, which is very different from how conventional processors operate and is the basis of the approach’s energy efficiency.
Large-scale neuromorphic systems are starting to come online. Intel’s Hala Point system combines 1152 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.5,6,7 Also, analogue neuromorphic computers, like the BrainScaleS system, that operate in continuous time, just as the brain does, are becoming available to the larger scientific community.8
There have also been breakthroughs in chips that carry out computation directly in memory.9 These are well-suited to implementing neuromorphic computing but also hold promise for reducing deep-learning AI’s energy consumption. That could lead to growing convergence between the two fields.10
However, the field is held back by a lack of shared software frameworks, which means that research remains fragmented and building neuromorphic models is far more complicated than it is conventional AI.11 A lack of efficient training algorithms also means performance still significantly lags behind deep learning, though there has been recent progress on that front.12,13 Insights from neuroscience could help close the gap,14,15 but faithfully replicating the brain’s function and structure in silicon remains a distant goal.
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:
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.
Rather than trying to create software and hardware that mimics the way the brain works, an emerging field of research seeks to coax nature’s most powerful computing technology — biological neural networks — into carrying out computations.
Improvements in the technology required to sustain and interface with organoids mean that they can now be easily integrated with conventional electronics in an ethical manner. They are routinely used for sensory functions in robotics, and interconnected networks of organoids are now able to carry out highly complex computations. Unconstrained by the body, brain organoids can be built at scales beyond anything possible in nature, allowing them to tackle novel computational problems. The rapidly improving capabilities of organoids provoke ethical work practices and discussion about the suppression of any development of consciousness and sentience.
This possibility has only just become tractable thanks to recent advances in organoids: simplified and miniaturised versions of organs created using stem-cell technology. Small conglomerations of human neurons have replicated some of the form and function of our brains.16 Now researchers are investigating whether these organoids could be used to create new hybrid computing technologies that combine biological and electronic components.17
Early experiments have shown that neural cultures can be taught to play video games, show features of reservoir computing or implement computer logic.18,19,20,21 If the technology can be scaled up it could have applications in AI, robotics and brain-machine interfaces.
Significant advances will be needed first, however, including engineering larger, more complex organoids, interfacing with them reliably and understanding how they learn and compute.22 There has been recent progress in more sophisticated “assembloids” that model several brain regions23,24 as well as boosting the diversity of cell types.25,26 New 3D electrode arrays are also providing more sophisticated interfaces27,28 and the extensive hardware required to sustain living cells is getting miniaturised.29
Dreams of living computers remain distant, however, and in the near term the technology will be a tool for brain research and biomedical testing. There are also ethical concerns around when, and whether, more sophisticated organoids could be thought of as conscious entities in their own right.30,31
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:
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.
Cellular computing can go beyond mimicking standard computing, though.33 A cell’s components can be reconfigured in response to external stimuli, and evolution allows populations of cells to adapt to changing environmental circumstances. They also function well in the presence of noise. There are multiple signal pathways within each cell, enabling massively parallel information processing.
This opens up the prospect of performing “whole-cell biocomputations” to solve challenges as varied as environmental remediation, drug discovery and medical diagnosis.34 There is increasing focus on harnessing networks of cells to carry out more sophisticated distributed computation35,36,37 and even incorporating viruses as an additional communication channel.38 Newly engineered plant-microbe communication channels open up the prospect of programming entire ecosystems.39
Neuromorphic computing is proving a useful framework for guiding adaptive cellular processes,40,41 though probabilistic computing could be another powerful conceptual lens.42 The field needs deeper engagement with theoretical computer science to give it a firmer conceptual footing, in particular to formalise how various synthetic-biology tools process information.43
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:
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.
These advantages have been known about for decades, but recent innovations, driven by investments from the telecoms industry, have started to make optical-computing devices practical. Breakthroughs in silicon photonics are making it possible to build sophisticated optical processors using the same technology as the existing chip industry.45
The most promising near-term application is in AI. Optical processors are very efficient at carrying out operations known as matrix multiplications, which are fundamental to deep-learning algorithms.46 Two startups targeting AI applications recently unveiled advanced processors that integrate both photonic and electronic components, suggesting the technology is on the cusp of commercial breakout.47,48
Optical approaches to quantum computing are also maturing fast, with breakthroughs in the mass manufacture of chips and the networking of large-scale devices.49,50 Photonic technology could also be used to build neuromorphic processors and analogue computing devices,51,5253,54 and the inherent noisiness of optical systems can be harnessed for tasks like probabilistic machine learning and optimisation.55,56
Catching up with decades of progress in silicon transistor technology remains an enormous engineering and financial challenge, though. In addition, photonic chips rely on wavelengths of light measured in micrometres, making it unlikely they could achieve the kind of miniaturisation found in silicon chips already boasting nanoscale features.
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:
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.
