PolyChord is a next generation data technology which has the ability to more fully explore a complex, challenging data landscape.
The kind of data you might get when a large array of sensors delivers data, and you want to get information out to give you insights into when predictive maintenance is needed. Or the kind of landscape obtained when you are looking at amino acids which are about to form a protein and you want a clearer view of how that protein might fold.
What are we using PolyChord for right now? Take a look at some of the examples below.
We currently have exciting new developments with Utilities infrastructure companies and also with telecommunications companies and we are keen to hear from other people in those areas.
Sensors in a Network
PolyChord is currently being contracted by the Ministry of Defence in the UK to use our unique technology in a project which optimises the placement of sensors in a Network. PolyChord’s high dimensional capabilities mean these techniques can optimise coverage for a whole city, state or country. PolyChord can also handle complex constraints such as “only place sensors on roofs”. Rather than keeping devices static in the planning phase, PolyChord resamples over the data space as new devices are added, allowing it to optimise over all possible positions of all devices efficiently to generate the best possible solution. This has key applications in the commercial world, including in 5G as constraints represented by buildings and other obstructions are taken into account. Optimising the placement of sensors using PolyChord reduces cost through less use of expensive hardware, increases the reliability of coverage and lowers your carbon footprint.
PolyChord are working with specialist sensor companies, using PolyChord’s ability to gain more useful information from the complex data landscape created by a set of sensors, to extract signal from noise, and then analyse and make predictive maintenance judgements in rail infrastructure. Working in a consortium of established companies in the sector, the end result is PolyTrack, which is revolutionary in using 5G to provide real-time decision making tools. By creating high-dimensional landscape maps of the data generated by these sensors, we are uncovering genuinely actionable insights for tram companies so that predictive maintenance is a reality rather than just an idea – we are then moving the approach into Light Rail, then on to mainstream Network Rail.
PolyChord is being used to predict the likelihood of pathogens occurring in food processing and storage. PolyChord offers more advanced model selection metrics than the W.H.O standard as we are able to account for more variables and to model atypical as well as normal growth. Working with leading biologists, our model has been developed using pathogens growth in low-moisture foods rather than lab strains, meaning our models are true to real-world scenarios. We have used actual lab work to back up our theoretical work, thanks to long collaborative input from Queen’s University Belfast pathogens expert Professor Seamus Fanning. Forward progress has been assisted thanks to a substantial Innovate UK grant. We are now turning our activities to focus on Listeria, until now a hard-to-track problem in food preparation and processing.
There are complex parameter optimisation challenges at the sophisticated side of 3D Printing – now more widely known as Additive Manufacturing. We are engaged with a leading company in the sector, improving their existing performance using PolyChord’s complex constrained optimisation capabilities to improve the production and selection of novel alloys for the aeronautics and space sectors.
Using PolyChord feature selection, we uncovered crucial information on what were the most salient parts of data sets, helping this innovative sensor technology company move closer towards FDA standards in POC Devices, sensing levels of particular substances in the blood without piercing the skin. Our feature selection ranks the features of your data set to tell you which is having the greatest impact on your output – allowing the “volume” to be turned up or down on certain metrics to achieve more accurate results.
Complex Industrial Processes
By creating a solution landscape, PolyChord allows for a better balance of trade-offs where a complex process has many objectives which a company or organisation might need to manage. If you turn up the volume on inputs A and D, how might it impact outputs P and Q? Our technology allows much more accurate insights and answers than were previously possible and is truly best in class at multi-objective/multivariate optimisation. We were recently invited to use it as part of a leading “Smart Factory” initiative and are now actively looking for new opportunities to apply this powerful technology. PolyChord copes easily with messy and incomplete data, without the need for a lengthy preparation process, which often bars well-established companies from engaging.
Could PolyChord deliver the kind of answers you could use in your business or organisation? We are open to discussion and happy to speak with you about the data you have and what you would like to get from it.
How does PolyChord achieve its results, and how was it invented in the Cambridge University Astrophysics Department to solve ten billion year large data problems? Take a look at the technical section. How do you work with organisations like mine? Check the FAQs
Please tell us some details about the nature of your enquiry