NVIDIA Ising AI models integrate with CUDA Q to automate quantum calibration and error correction decoding for large scale computing
NVIDIA Ising aims to achieve its goal of using AI technology to control quantum systems. The company introduced its Ising open model family to create a connection between experimental qubits and large scale quantum computing. The company will use artificial intelligence as its main control system to control quantum hardware fluctuations, which will enable the industry to transition from theoretical research to fast supercomputing.
The current quantum system needs better processor calibration and error correction because existing techniques take too long and produce too many errors. The Ising framework directly addresses these bottlenecks. NVIDIA claims that the models boost error correction decoding speed by 2.5 times while delivering three times better accuracy compared to the current industry standard known as pyMatching. Jensen Huang, the CEO of NVIDIA, described this transformation as making AI the "operating system of quantum machines" which delivers stability for hybrid quantum GPU systems.
The Ising family is divided into two primary technical functionalities to assist researchers. The Calibration model uses a vision language interface to understand live quantum processor measurements which reduces calibration times from multiple days to just hours. The Decoding model uses 3D convolutional neural networks to conduct real time quantum error correction through its two available modes which offer either speed or accuracy optimization. The team will receive these tools together with a complete cookbook containing workflows and NIM microservices, which lets them enhance their local operations while keeping their data secure.
Most of the current framework users come from research organizations and business companies that operate throughout the world. Ising Calibration helps organizations like Fermi National Accelerator Laboratory and Harvard’s School of Engineering and Applied Sciences and the U.K. National Physical Laboratory to operate their research facilities. The Decoding components are being used by Cornell University and Sandia National Laboratories and various international research collaborators, which shows how NVIDIA software is becoming a common standard in quantum research.
The initiative expands NVIDIA's open model portfolio by introducing a new model. The company is building a complete operating system for hybrid quantum and classical computation through the integration of Ising with both CUDA Q and NVQLink systems. The quantum computing industry will reach a market value of $11 billion by 2030, yet AI based automated error correction will function as the essential engineering component needed to resolve current quantum processor stability challenges.
