Probabilistic inference
Use hardware that deliberately produces variable outputs for tasks involving uncertainty or many possible answers.
PrototypePost-Digital Primer 09
Other and emerging domains
A disciplined holding category for new mechanisms that cross boundaries or have not yet earned a permanent domain.
Why it matters
Important technologies often appear before the market agrees what to call them. This category preserves credible but hard-to-classify signals without promoting every novel laboratory device into a supposed new industry.
What it is
This category tracks approaches that do not yet fit the nine established domains. Probabilistic bits deliberately fluctuate between states. Ising or oscillator machines turn an optimisation problem into a network that searches for a stable arrangement. Thermodynamic circuits use controlled heat-driven fluctuations. Reversible logic tries to avoid discarding information, while magnonics processes signals using waves in magnetic materials. Mechanical intelligence and embodied computation place some control in the physical design of a machine rather than in software alone.
What it is not
It is not a bucket for weak news. An item belongs only when the computing mechanism is materially different, the source offers technical evidence, and forcing it into another family would hide the main idea. A recurring cluster graduates to its own explainer.
Telecoms read-acrossImportant new architectures may not fit an established label at first. A controlled miscellaneous section preserves early signals without pretending that every laboratory curiosity is a new industry.
How it works
01Admit
Identify a distinct information-processing mechanism and a substantive technical source.
02Compare
Test whether an existing family already explains the claimed advantage.
03Track
Follow repeatability, independent groups, software and a plausible workload.
04Graduate
Create a permanent family when evidence and activity become coherent enough.
Use cases
Potential uses.
Status labels show use-case maturity.
Ising optimisation
Represent a difficult allocation problem as a network that settles towards a low-cost arrangement.
Research to early systemsThermodynamic computation
Use controlled heat-driven fluctuations to perform selected calculations with low energy.
Concept and simulationMagnonic processing
Use collective waves in magnetic materials to carry, filter or transform microwave-frequency information.
Laboratory researchTelecom relevance
Possible telecom roles.
Network optimisation
Probabilistic and Ising machines may compete with quantum and classical solvers on allocation problems.
Edge robotics
Embodied mechanics could reduce sensing, control and communications loads in constrained machines.
Semiconductor strategy
Watch which mechanisms reach standard processes rather than requiring exotic fabrication.
Research scouting
Use this category to spot converging clusters before they acquire stable terminology.
Reality check
Pre-requisites and proof-points needed
These questions test whether the technology can become a practical, supportable system.
- 01Is the mechanism distinct or merely rebranded analog compute?
- 02Has more than one group reproduced it?
- 03Is there a workload with a measurable advantage?
- 04Should this cluster now graduate to a permanent domain?
Related signal coverageSignals Unclassified and cross-category signals
Browse the signal archive →Back to all explainers ↑