TORIDION describes itself as a quantum computing company. But our goal was never to compete in a quantum race. True to our problem solving roots out of Visicom Scientific Software's R+D efforts, TORIDION instead started with an actual problem and sought practical solutions.
Beginning with data compression, and ultimately the development of Machine Learning, we recognised and subsequently demonstrated that the core TORIDION algorithms were capable of solving specific problems currently intractable for classical approaches.
Our quantum heritage comes from our technology stack which begins with the quantum inspired algorithms. The core TORIDION algorithms favour Machine Learning and Grover type search applications.
For a widening range of real world problems, our algorithms and simulation demonstrates massive performance gains whilst running on classical processor architecture. More than this, they also represent a ground breaking speedup in terms of solving certain classically intractable problems. This is in the main, possible because of the quantum heritage of the algorithms employed.
Beyond the obvious and demonstrated computational advantages of our algorithms, our differentiator lies in how the algorithms were designed. Built from day one to lever quantum analog processors being developed by TORIDION which will drastically enhance the whole system.
Our algorithms implement multiple mathematical superpositions and conversions that are purposefully designed to encode a multitude of classically intractable steps into a single virtual computational unit. This computational unit (or the workload) is not easy (if even possible) to solve on a classical computer – although our algorithms running on simulator, even burdened with these massive overheads still outperforms popular classical approaches for a number of very useful problems.
The workload encoded by the algorithm is (by design) solvable instantly by a number of quantum analog approaches – approaches that do not require superconducting qubits or other complex and fragile hardware. Because the virtualised workload requires many millions of classical steps to solve, when we externalise the workload and solve it using quantum analog co-processing, we effectively increase performance many many orders of magnitude.
The perfection of Quantum Analog computational co-processors will allow our algorithms to simply offload calculations that currently require millions to trillions of classical cycles to a single low energy processor that will solve the equation natively and with zero latency.
Much like any quantum computing approach, utilising the TORIDION becomes about effective data preparation. Learning how to prepare problem inputs in a way that encodes them into the workload container so it can be solved by the simulator or co-processor. We continue to develop new and novel training algorithms.
Recognising that useful universal quantum computers may be 10+ years from fruition, our approach is both legitimate and valuable as it allows the delivery of quantum value now, whilst using an easy to integrate technology stack that will only improve as quantum analog processing hardware becomes available.