Massive Analytic's Quantum Leap

With the announcement from Theresa May this week that the UK government is pledging £153 million to further develop Quantum Computing, it is the hot topic in UK tech right now. And it’s not just here in the UK that Quantum Computing is making big waves, across the pond the US government is also currently fine-tuning it’s Quantum Computing policy. Many are claiming that Quantum Computing will be the next industrial revolution, and it’s hard to argue that the potential isn’t there for a seismic shift in society if the advances take off.

But where does Massive Analytic fit into all this? Well Quantum Computing is an area that in the past few months has become keenly relevant to us as, we’ve just kicked off a new project which catapults us into this space - the Quantum Point Cloud (QPC) Representation Feasibility Study.

Partnering with the National Physical Laboratory (NPL), this project is a milestone for us because it’s our first step into the quantum space, a new market which has the potential to revolutionise not just AI and data science but also society as a whole. Funded by Innovate UK, this project seeks to establish whether fused multi-modal sensor data can be represented as a QPC, if successful it will be a giant leap forward for processing and later analysing huge fused data sets. Our hope is that a successful project will provide the foundation for a new set of applications, commercialised through Massive Analytic, that will drastically simplify the process and reduce the cost of gaining insights from fused multi-model data – building on top of our present capabilities in those areas.

I won’t go into the technicalities of the project too deeply but just to lay out the basics, a Classical Point Cloud (CPC) is a set of data points in a three-dimensional coordinate system that represents the external surface of an object. CPCs can be generated through hardware sensors such as LiDAR, a set of RGB images and from thermal imaging to name just a few. To give an example of what shape a CPC might take, it could be a representation of an urban environment generated by LiDAR that is then converted into a 3D model. However visualising a CPC is extremely difficult due to the processing requirements, because in order to record the surface clearly, a CPC usually has millions, or even hundreds of millions, of points (surfaces in the real world are rarely perfectly smooth after-all), this means conventional computers often crash when operating a CPC of multi-modal sensor data.

This is where Quantum Computing comes in, classical computers use bits to store information and execute commands – these are binary and can be either 0 or 1, this is because classical computers as the name implies are limited by classical physics. However in Quantum Computing, a bit can be both 0 and 1 at the same time. This is because the laws of quantum physics allow electrons to be in multiple places at one time. This superposition means that “quantum bits” or “qubits” not only store way more information than a classical bit but that quantum computers can process information on a completely different level to current computers. Which would be pretty handy for processing huge point clouds of fused multimodal data!

The thesis underpinning this project is that quantum computers can help users to efficiently solve the multi-modal sensor CPC processing problem. Here we can draw parallels with Quantum Image Processing (QIP) which is primarily devoted to using Quantum Computing and Quantum Information Processing to create and work with quantum images. A Quantum Image is an image where the information is encoded in quantum-mechanical systems e.g. qubits, instead of classical ones. In this area quantum computation is becoming an important and effective tool in overcoming the high real-time computational requirements of classical digital image processing, of which multi-modal sensor fusion is just one example. It’s easy to understand why research is turning to QIP when millions of photo’s are being shared every single day. In the same way QIP is seen as way to manage resources in a society that heavily shares images, QPC could be the solution to the growing challenge posed by the Internet of Things (IoT) and leveraging the insights from the growing volume of sensor data being produced by today's digital world.

A positive outcome or deliverable from this feasibility study will not serve a ‘market niche.’ It’s much grander than that. The potential applications will underpin the future of humankind. The use of fused multi-sensory data is increasingly being viewed as a way in which state and private sector organisations can do ‘public good’ by improving service and product delivery at a lower cost. By doing more with less. We believe that the importance of QPC numerical representation is a first order priority in advancing technology that is not well understood. In this respect, pin-pointing ‘market niches’ or applications which can benefit from a positive outcome of this project is a difficult task. The list would be nearly endless. However there are several other projects that we're working on that will directly benefit from this research - but more on those at a later date.

Massive Analytic is only at the beginning of it’s quantum journey, but the signs are hopeful that QPC’s could pave the way for advances in all manner of areas, from autonomous vehicles, to precision medicine, to agriculture, both in terms of new ways of processing data and developing new better quantum algorithms – the possibilities are exciting and massive.