With the responsibility to tool up core software application development practices for the agentic AI revolution, a similarly significant gravitational force may soon be felt as a result of quantum computing more broadly. Within the arena of quantum, photonic (optical) quantum computers in particular are an emergent branch of the field that encodes quantum information in particles of light.
To draw the distinction for the uninitiated, where traditional computing systems rely on established binary logic, quantum machines work on a different principle. Their qubits can exist in a superposition of both 0 and 1 at the same time, only resolving to a definite 0 or 1 when they are measured.
Going further then, optical quantum computers are machines that use individual photons as their qubits, manipulating them through circuits of beam splitters and phase shifters (the mechanical gearing systems of photonics, if you will) etched onto chips.
How To Program Optical Quantum Apps
Developers fresh out of college with a gleaming set of Java, Python, Rust and Go skills may at this point be questioning where the toolkits will sit in the space, especially given all the flux right now between platform engineering and agentic copilot services.
If it’s going to take a big player to serve this space, then NTT, Inc. execs think the company is up for the job. The company this month announced that it is advancing research and development of software technologies to expand the use of optical quantum computers and enable their future practical application.
Specifically, NTT is working on system software that includes a quantum application software development kit (SDK), which is hoped to make it easier to develop machine learning applications. Alongside a compiler designed for the characteristics of optical quantum computers, there are also optimization controls and an execution platform for hybrid quantum-classical processing.
Digging deeper, this work also sees the development of technologies for handling “noise” on actual hardware (unwanted interference, such as photon loss, that corrupts fragile quantum calculations) and for future fault-tolerant quantum computers.
What ‘Shape’ Is This SDK?
In terms of form and function, programmers will naturally want to know what kind of shape this SDK is (by which we mean what levels of abstraction and integration are on offer) – and it appears that NTT has confirmed coders will be able to build applications simply by describing the machine learning processing, without directly designing or writing quantum circuits.
The SDK internally constructs the quantum circuits required for machine learning processing. This reduces the burden of designing and writing quantum circuits and provides an environment in which users can focus on application development and validation. NTT used the SDK to develop a machine learning application, connected it to the optical quantum computer MoQuren via an SDK provided by [Japanese optical quantum specialist] OptQC Corp and verified end-to-end execution from the application to the actual hardware.
“Going forward, NTT will expand the functions and supported fields of its quantum application SDK, advance software technologies that draw out the full performance of optical quantum computers, publish technical information, and co-develop and validate quantum applications with partner companies. Through these efforts, NTT will develop use cases in industries such as finance, manufacturing and drug discovery,” noted the company, in a press statement.
The Quantum Programming Challenge
NTT reminds us that bringing optical quantum computers into practical use requires not only better quantum hardware performance but also a software environment that allows developers to build and run applications that use optical quantum computers.
“However, using a quantum computer can require specialized knowledge, such as how to design and write quantum circuits and how to execute them on specific hardware. Technologies are also needed to make computation more efficient and results more reliable, including optimizing processing for the characteristics of quantum computers, coordinating quantum and classical computation, and handling the noise that occurs on actual hardware,” stated NTT.
The SDK provides Application Programming Interfaces (APIs) for describing machine learning processing and internally constructs the quantum circuits required for that processing. As a result, users can develop machine learning applications without directly designing or writing quantum circuits, which reduces the specialized knowledge and effort required for quantum circuit design.
Processing is carried out in the following steps:
1. Writing the application: Using APIs for machine learning, users describe the machine learning processing without directly designing or writing quantum circuits.
2. Constructing quantum circuits and connecting to the hardware: NTT’s SDK internally constructs the quantum circuits required for the machine learning processing and connects to the optical quantum computer MoQuren via the SDK provided by OptQC.
3. Running computations on MoQuren and using the results: Optical quantum computations are executed on MoQuren, and the results are used by the machine learning application.
NTT will expand the functions and supported fields of its quantum application SDK, targeting application areas such as machine learning, optimization, and simulation. The goal is an environment in which a wide range of applications can use optical quantum computers without users having to be aware of designing or writing quantum circuits.
To draw out the full performance of optical quantum computers, NTT will also advance research and development of a compiler that converts and optimizes quantum circuits, an execution platform that enables hybrid quantum-classical processing and efficient use of computing resources, and technologies that reduce the impact of noise on actual hardware. In addition, with a view to future scale-up and support for fault-tolerant quantum computers, NTT will enhance software technologies for error correction and for improving overall system performance.
Developers, Don’t Panic
If this all still sounds a little cutting-edge, NTT promises it will share technical information on software for optical quantum computers through technical blogs and other channels and promote adoption among developers. At the same time, based on the challenges faced in industries such as finance, manufacturing, and drug discovery, NTT will co-develop and validate quantum applications with partner companies to develop use cases that lead to practical applications.
Developers, you’ll still need your Java and Python prowess – after all, the likelihood is that much of this new engine power will get packaged into commoditized boosters that don’t affect the programmer’s daily workflow, but do provide superhero turbo charges at the back end – so go forth on the command line, but be aware of infrastructural change.
