Together with Silicon Quantum Computing (SQC), we’ve taken Quantum Computing out of the lab and into the field to solve complex network challenges.
Over the past 12-months a joint Telstra-SQC team has explored how quantum-enhanced machine learning could enhance Telstra’s response to a complex challenge in connectivity – predictive analytics.
Currently, we use a combination of machine learning and AI to help predict network performance and detect changes in network patterns. Such systems analyse metrics across the board - such as latency and bandwidth to predict potential variances.
These analytics technologies help us to make decisions in where we supply bandwidth across our network as well as let us know if there are any faults in the network that need addressing – often before there’s even any customer impact. They can also help deliver greater personalised services to customers, including through things like dynamic bandwidth upgrades that can respond to changes in demand - in real-time.
Our joint team tested SQC’s quantum-enhanced machine learning system called Watermelon, a quantum reservoir that generates quantum features that can be used in an AI deep learning model.
SQC’s quantum experts and Telstra’s engineers wanted to understand if features generated by the quantum reservoir could be used to forecast network metrics - and compare the performance to a recently developed deep learning model.
While not live data, the test was conducted on real data and on SQC’s locally developed silicon-based quantum processor – not an off the shelf quantum cloud product.
The result?
Compared to the existing deep learning model - which required weeks of training and fine-tuning parameters - the team were able to train and fine-tune the reservoir’s parameters in just a few days, achieving comparable accuracy.
For us, we see this as huge breakthrough demonstration on how quantum-enhanced AI could realistically be used within the telco industry and on the Telstra network.
In this example, think of reservoirs like a magical pond that reacts to network data. You throw in data and watch the ripples. The shape of the ripples tells you something about how the network data is all related.
In computing, a reservoir is a complex system that reacts to input data in a rich, dynamic way. You don’t change the pond itself - you just observe how it reacts.
Telco networks like ours generate huge streams of data. Think call logs, signal strength, latency, etc. Most of the time, this data follows predictable patterns - and that’s good because consistent patterns mean we can make future predictions of what’s normal.
To take this back to our pond - a quantum reservoir uses the properties of quantum to create even richer, more complex ripples as the data is passed through it. The quantum ripple effect sees data exploded into many dimensions, which means a more detailed view of it.
These quantum ripples can represent predicable patterns we usually see while identifying other less-clear patterns in data that classical systems might miss when trying to the same thing. These subtle patterns are important because they can be used to tell us that something unusual is about to happen or is, in fact, already happening.
Think of it like upgrading your pond to a magical pond that reacts in ways normal physics can’t explain — giving you deeper insights.
Unlike traditional deep learning models, which demand significant GPU (Graphics Processing Units) resources and manual fine-tuning, the quantum reservoir required minimal setup and operated efficiently without the need for GPUs.
In the past, to train a classical deep learning model to do this type of analysis would take weeks and much compute power. In this trial, to bring Watermelon into the same performance level took us just a few days – without accuracy loss.
Quantum computing is a promising frontier we’re continuing to explore for a number of reasons. With the current global demand for GPUs, and sustainability goals, it has the future potential to identify unusual network behaviour without, those GPU demands while doing so faster than ever.
Telstra is an investor in SQC to drive innovation, build Australian-made digital infrastructure, and secure access to leading quantum expertise.
As Telstra and SQC continue to collaborate, the results from this trial lay the foundation for broader investigation of quantum technology in digital infrastructure and broader real-world industry applications.
The collaboration shows how Australian industries and homegrown innovation can work together to shape the nation’s digital future.
