The problem:
Every data value obtained from an empirical source, such as from a real-world event, has some amount of measurement uncertainty associated with it. Despite this reality, computers today do not intrinsically track how uncertainties in their inputs affect their internal calculations and eventual outputs. This inability to track uncertainties in data processed leaves existing computing systems unable to determine when their results are meaningless. This can have severe consequences when data-driven computing systems are used in applications such as quantitative finance, industrial control automation and autonomous vehicles (e.g., self-driving cars, unmanned aerial vehicles, and unmanned underwater/surface vehicles).
The solution:
Signaloid’s technology can be designed into computing hardware such as microprocessors, GPUs, and FPGAs, as well as measurement hardware such as sensors. Our technology allows computing hardware to track empirical uncertainties associated with the data processed, and allows measurement systems to encode uncertainty information into every individual data value they generate.
Our technologies are designed to be integrated seamlessly into hardware, without requiring any changes in existing software, or to other parts of existing systems. Because our technology is more than two orders of magnitude faster than existing alternatives, it makes possible, for the first time, data-driven systems that can continuously track the quality and accuracy of their data-driven decisions.
The global semiconductor market reached over $400 billion in 2020, with the global sensor market valued at around $200 billion and the market for logic at around $100 billion. Of this, the AI-related semiconductor market was valued at $6 billion and has since been rapidly growing, expected to reach $30 billion this year. PwC recently highlighted that to maximise its potential, future digital semiconductor systems will need to be able to provide users with timely information on their operations. This is something Signaloid’s technology will be fundamental to, and we intend to make possible in the near future. We currently have one version of our technology already deployed and being used by paying customers, and over the next 12–18 months, our goal is to broaden the set of customers using the current form of our technology, as well as to expand the form in which we make the technology available, allowing us to break into new markets.
Traction:
- Signaloid’s technology is already being used as a key part of a £2 million Innovate UK Demeter project (2021–2023).
- The underlying research that led to Signaloid was recently selected as one of the Top Picks in Computer Architecture Research (2022).