Quantum Intelligence with Light:

Building Quantum Large Language Models using Photonic Analog Quantum Computing

Why Photonic?

Most of the physical implementations of quantum computing require temperature control close to absolute zero Kelvin. A Quantum Processing Unit using Quantum Optics was developed by Xanadu.

Advantages:

Compatible with the existing communications infrastructure.
Operates at room temperature.
Higher dimensional computational space.
Easy to network and multiplex.
Low cost for mass production.
Mountable on smartphones, laptops, and edge devices.

Why Analog?

Nature is continuous, not binary.

Nature isn't classical, dammit, and if you want to make a simulation of nature, you'd better make it quantum mechanical, and by golly it's a wonderful problem, because it doesn't look so easy.

Richard Feynman

Theoretical physicist

We are using the binary system in digital computing because of the ON and OFF switches of transistors. It is a hardware constraint that need not be dragged into the quantum world.

Quantum systems are continuous. In quantum devices for computing, we are free to use the continuous variable logic implemented in Analog Quantum Computing.

Computational Space: Classical to Quantum

Moving from the computational space consisting of 0 and 1 to the space of infinitely many points gives a huge advantage of encoding and processing information.

The superposition property of quantum states gives inherent parallelism in quantum computing. With the higher dimensional computational space in Analog Quantum Computing, a higher level of parallelism is achieved.

Classical

Qubit-based

Continuous Variable

Wagner function of |0⟩

Wagner function of |1⟩

Wagner function of |2⟩

Wagner function of |3⟩

Quantum Large Language Model

The building blocks of LLMs are transformers. By replacing the Feedforward blocks with Quantum Neural Networks in transformers, we develop Quantum LLMs.

Continuous Variable Quantum Neural Network

Quantum Transformer

Quantum LLM

Related Papers

Quantum computing overview: discrete vs. continuous variable models

In this Near Intermediate-Scale Quantum era, there are two types of near-term quantum devices available on cloud: superconducting quantum processing units (QPUs) based on the discrete variable model and linear optics (photonics) QPUs based on the continuous variable (CV) model. Quantum computation in the discrete variable model is performed in a finite dimensional quantum state space and the CV model in an infinite dimensional space. In implementing quantum algorithms, the CV model offers more quantum gates that are not available in the discrete variable model. CV-based photonic quantum computers provide additional flexibility of controlling the length of the output vectors...

Quantum circuits with many photons on a programmable nanophotonic chip

Growing interest in quantum computing for practical applications has led to a surge in the availability of programmable machines for executing quantum algorithms. Present day photonic quantum computers have been...

Continuous-variable quantum neural networks

We introduce a general method for building neural networks on quantum computers. The quantum neural network is a variational quantum circuit built in the continuous-variable (CV) architecture, which encodes quantum...

Continuous Variable Quantum MNIST Classifiers —Classical-Quantum Hybrid Quantum Neural Networks

In this paper, classical and continuous variable (CV) quantum neural network hybrid multi-classifiers are presented using the MNIST dataset. Currently available classifiers can classify only up to two classes. The...