Telstra & SQC trial quantum AI to speed network analytics

Telstra has announced the results of a trial partnership with Silicon Quantum Computing (SQC) to develop quantum-enabled infrastructure to improve network predictive intelligence.

Over a 12-month period, teams from Telstra and SQC explored the use of SQC's quantum reservoir technology, named Watermelon, to investigate how quantum machine learning could enhance Telstra's predictive analytics capability for its telecommunications network.

Quantum reservoir testing

The trial aimed to address key challenges in connectivity, particularly in using predictive analytics to improve customer experience. Predictive network analytics play a significant role in pre-empting and resolving network issues, as well as supporting services like dynamic bandwidth upgrades that respond in real time to customer demand.

Currently, Telstra relies on a combination of machine learning and artificial intelligence to analyse network metrics, including latency and bandwidth, to detect and predict changes in network performance. These systems enable proactive monitoring and adjustment of network resources, deployment of technicians, and automated responses and actions often implemented before the impact reaches the customer.

SQC's quantum reservoir, Watermelon, was tested and evaluated by a combined Telstra and SQC engineering team. The system generates quantum features that can be incorporated into artificial intelligence models. The evaluation focused on two primary objectives: determining whether quantum reservoir features could successfully forecast network metrics, and comparing their performance to Telstra's existing deep learning models.

Performance outcomes

According to the companies, the results showed that training and fine-tuning the quantum reservoir took several days, compared to the weeks required for the current deep learning models. Despite the reduced training time, accuracy remained comparable to the traditional deep learning systems.

The quantum model also functioned without relying on the computing power provided by graphics processing units (GPUs), which are commonly needed for deep learning. With resource use and the cost of high-performance AI infrastructure becoming increasingly relevant operational concerns, the trial demonstrates the potential value of more efficient quantum approaches.

Industry perspectives

Shailin Sehgal, Telstra's Group Executive of Global Networks and Technology, said:

"We're constantly looking ahead to technologies that can help us create smarter connectivity experiences for our customers - from increased personalisation to issue prevention. Quantum computing is a promising frontier we're exploring. Working with SQC allows us to research the real-world potential of quantum systems in a uniquely Australian context

"This trial shows how quantum capabilities could complement our existing systems and technology to deliver faster insights and better outcomes for our customers. The collaboration, and Telstra's relationship with SQC, shows how Australian industries and homegrown innovation can work together to shape the nation's digital future."

Michelle Simmons, Chief Executive Officer of Silicon Quantum Computing, also reflected on the trial's broader significance for commercial quantum adoption:

"This is an exciting and important step forward in commercial adoption of quantum technologies. The collaboration with Telstra allowed us to test our quantum reservoir system, Watermelon, in a real-world telecommunications context - something few quantum companies have achieved. Watermelon's quantum feature generation helps to reveal complex relationships within classical data, while dramatically reducing training time."

"We've always believed that the key to unlocking quantum's full potential lies in building systems with atomic precision and purity. This partnership shows how quantum processors have moved beyond theory and into practical, scalable solutions that enhance Australia's digital infrastructure."

Quantum computing background

Quantum computing applies the fundamental laws of quantum mechanics to computation, notably through the use of qubits, which can exist in multiple states at once through superposition, and can exploit entanglement to process information in parallel. These properties allow quantum computers to tackle problems that are infeasible for classical computers to address exhaustively.

Quantum reservoirs leverage the dynamic and nonlinear properties of quantum systems to process time series data, which is common in network analytics. They can offer greater resilience to noisy or sparse data compared to traditional deep learning methods. They are of particular interest in large-scale systems with recurring patterns and multiple data inputs, such as forecasting capacity, dynamic workload assignment, and network assurance functions.

Through this trial program, Telstra and SQC aim to lay a foundation for expanded investigation into the application of quantum technology in digital infrastructure, with the potential for broader real-world impact across the telecommunications and technology sectors.