QC Ware, a startup in software and services for quantum computing, announced a significant breakthrough in quantum machine learning (QML) that increases QML accuracy and speeds up the industry timeline for practical QML applications on near-term quantum computers.
QC Ware’s algorithms researchers have discovered how classical data can be loaded onto quantum hardware efficiently and how distance estimations can be performed quantumly. These new capabilities enabled by Data Loaders are now available in the latest release of QC Ware’s Forge™ cloud services platform, an integrated environment to build, edit, and implement quantum algorithms on quantum hardware and simulators.
Apart from the Forge Data Loaders, the latest release of Forge includes tools for GPU acceleration, which allows algorithms testing to be completed in seconds versus hours, and turnkey algorithms implementations on a choice of simulators and quantum hardware. Simulations are executed on CPUs and Nvidia GPUs on AWS. Quantum hardware integrations include D-Wave Systems, and IonQ and Rigetti architectures through Amazon Braket.
Forge offers two types of data loaders: the Forge Parallel Data Loader and the Forge Optimized Data Loader, which optimally transform classical data to quantum states to be readily used in machine learning applications. Additionally, QC Ware is introducing optimized Distance Estimation algorithms that allow for powerful quantum classification and clustering applications.
These capabilities were considered major challenges for QML algorithms. Most research papers from academia, government, and industry assume the availability of Quantum Random Access Memory (QRAM), the quantum equivalent of classical RAM, to load data on quantum computers. However, very few researchers and vendors have worked on QRAM, and the few proposals around it come with very significant hardware requirements in qubit count and circuit depth. The Forge Data Loaders provide a powerful and near-term alternative to QRAM.