Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:2504.04009 · Optics Express

Engineering nonlinear activation functions for all-optical neural networks via quantum interference

Ruben Canora, Xinsen Xu, Ziqi Niu, Hadiseh Alaeian, Shengwang Du

Published Oct 24, 2025• 7 citations
arXivHTMLPDF

Abstract

All-optical neural networks (AONNs) promise transformative gains in speed and energy efficiency for artificial intelligence (AI) by leveraging light's intrinsic parallelism and wave nature. However, their scalability has been fundamentally limited by the high power requirements of conventional nonlinear optical elements. Here, we present a low-power nonlinear activation scheme based on a three-level quantum system driven by dual laser fields. This platform introduces a two-channel nonlinear activation matrix with self- and cross-nonlinearities, enabling true multi-input, multi-output optical processing. The system supports tunable activation behaviors, including sigmoid and ReLU functions, at ultralow power levels (17 μ W per neuron). We validate our approach through theoretical modeling and experimental demonstration in rubidium vapor cells, showing the feasibility of scaling to deep AONNs with millions of neurons operating under 20 W of total optical power. Crucially, we demonstrate the all-optical generation of gradient-like signals with backpropagation, paving the way for all-optical training. These results mark a significant advance toward scalable, high-speed, and energy-efficient optical AI hardware.

DOI
10.1364/oe.578666

Related papers

Optical Neural Networks from Coherent Transient Dynamics in Waveguide QED

Jiande Cao, Yexiong Zeng, Franco Nori +1 more

Optical neural networks promise ultrafast, low-energy information processing by performing computation directly with photons. Current implementations, however, are largely restricted to steady-state operation and rely on high-latency electro-optical conversion for nonlinear activation. To address these limitations, we propose an all-optical fully connected neural network architecture in which the basic neuronal functions are realized by coherent transient quantum dynamics. Within this framework, phase-tunable nonlocal interference in a giant cavity implements programmable synaptic weights; an integrator operating in the bad cavity regime performs temporal summation by coherently combining sequential wavepackets; and transient Rabi dynamics of a driven two-level system provide nonlinear activation. Full-physics simulations demonstrate high classification accuracy on MNIST and colored-object recognition tasks. These results eliminate the optoelectronic activation bottleneck, reduce latency, and establish transient light-matter dynamics as a native physical resource for high-dimensional nonlinear information processing, paving the way toward fully optical neuromorphic computing.

quant-pharXiv:2605.17752•May 19, 2026•0 citations

Quantum optical neural networks using atom-cavity interactions to provide all-optical nonlinearity

Chuanzhou Zhu, Tianyu Wang, Peter L. McMahon +1 more

Optical neural networks (ONNs) have been developed to enhance processing speed and energy efficiency in machine learning by leveraging optical devices for nonlinear activation and establishing connections among neurons. In this work, we propose a quantum optical neural network (QONN) that utilizes atom-cavity neurons with controllable photon absorption and emission. These quantum neurons are designed to replace the electronic components in ONNs, which typically introduce delays and substantial energy consumption during nonlinear activation. To evaluate the performance of the QONN, we apply it to the MNIST digit classification task, considering the effects of photon absorption duration, random atom-cavity detuning, and stochastic photon loss. Additionally, we introduce a convolutional QONN to facilitate a real-world satellite image classification (SAT-6) task. Due to its compact hardware and low power consumption, the QONN offers a promising solution for real-time satellite sensing, reducing communication bandwidth with ground stations and thereby enhancing data security.

quant-pharXiv:2511.06167•Physica Scripta•Nov 11, 2025•3 citations

Universal logical quantum photonic neural network processor via cavity-assisted interactions

Jasvith Raj Basani, Murphy Yuezhen Niu, Edo Waks

Encoding quantum information within bosonic modes offers a promising direction for hardware-efficient and fault-tolerant quantum information processing. However, achieving high-fidelity universal control over bosonic encodings using native photonic hardware remains a significant challenge. We establish a quantum control framework to prepare and perform universal logical operations on arbitrary multimode multi-photon states using a quantum photonic neural network. Central to our approach is the optical nonlinearity, which is realized through strong light-matter interaction with a three-level Λ atomic system. The dynamics of this passive interaction are asymptotically confined to the single-mode subspace, enabling the construction of deterministic entangling gates and overcoming limitations faced by many nonlinear optical mechanisms. Using this nonlinearity as the element-wise activation function, we show that the proposed architecture is able to deterministically prepare a wide array of multimode multi-photon states, including essential resource states. We demonstrate universal code-agnostic control of bosonic encodings by preparing and performing logical operations on symmetry-protected error-correcting codes. Our architecture is not constrained by symmetries imposed by evolution under a system Hamiltonian such as purely χ (2) and χ (3) processes, and is naturally suited to implement non-transversal gates on photonic logical qubits. Additionally, we propose an error-correction scheme based on non-demolition measurements that is facilitated by the optical nonlinearity as a building block. Our results pave the way for near-term quantum photonic processors that enable error-corrected quantum computation, and can be achieved using present-day integrated photonic hardware.

arXiv:2410.02088•npj Quantum Information•Aug 20, 2025•7 citations

Hardware-inspired Continuous Variables Quantum Optical Neural Networks

Todor Krasimirov-Ivanov, Alba Cervera-Lierta, Paolo Stornati +1 more

Continuous-variables (CV) quantum optics is a natural formalism for neural networks (NNs) due to its ability to reproduce the information processing of such trainable interconnected systems. In quantum optics, Gaussian operators induce affine mappings on the quadratures of optical modes while non-Gaussian resources (the challenging piece for physical implementation) originate the nonlinear effects, unlocking quantum analogs of an artificial neuron. This work presents a novel experimentally-feasible framework for continuous-variable quantum optical neural networks (QONNs) developed with available photonic components: coherent states as input encoding, a general Gaussian transformation followed by multi-mode photon subtractions as the processing layer, and homodyne detection as outputs readout. The closed-form expressions of such architecture are derived demonstrating the family of adaptive activations and the quantum-optical neurons that emerge from the amount of photon-subtracted modes, proving that the proposed design satisfies the Universal Approximation Theorem within a single layer. To classically simulate the QONN training, the high-performance QuaNNTO library has been developed based on Wick-Isserlis expansion and Bogoliubov transformations, allowing multi-layer exact expectation values of non-Gaussian states without truncating the infinite-dimensional Hilbert space. Experiments on supervised learning and state-preparation tasks show balanced-resource efficiency with strong expressivity and generalization capabilities, illustrating the potential of the architecture for scalable photonic quantum machine learning and for quantum applications such as complex non-Gaussian gate synthesis.

quant-pharXiv:2512.05204•Dec 8, 2025•1 citation

Quantum-Inspired Neural Networks

R. Raihana Parveen, N. S. Kalyan Chakravarthy, S. Jafar Ali Ibrahim +3 more

Two strong transformative fields in contemporary artificial intelligence (AI) are quantum computing and deep learning. Neuro-nanophotonic systems also introduce quantum-inspired innovations that enhance the computing power of neural networks. Quantum-Inspired Neural Networks (QINNs) have shown promise as a potential alternative to the classical deep learning models, particularly when processing high-dimensional data and solving complex problems efficiently. QINNs leverage the attributes of quantum mechanics to enable efficient deep learning methodologies, achieving superior performance in AI-based applications. In this research, we present a QINN framework that synergizes the power of Quantum Variational Circuits (QVCs) and Quantum Natural Gradient Descent (QNGD) to improve learning efficiency of learning. Experimental results show that QINNs can achieve 97% accuracy, which is 4.5% improvement over the classical models, and computational time and energy consumption can be reduced by 45% and 53%, respectively. Furthermore, QINNs exhibited greater robustness to adversarial noise, making them better candidates for healthcare, cybersecurity, and large-scale AI applications. The results illuminate the game-changing ability of QINNs with respect to scalable, cross-computational and highly resilient AI architectures.

CRC Press eBooks•Mar 31, 2026•0 citations

Towards a Real Quantum Neuron

Wei Hu

Google’s AlphaGo represents the impressive performance of deep learning and the backbone of deep learning is the workhorse of highly versatile neural networks. Each network is made up of layers of interconnected neurons and the nonlinear activation function inside each neuron is one of the key factors that account for the unprecedented achievement of deep learning. Learning how to create quantum neural networks has been a long time pursuit since 1990’s from many researchers, unfortunately without much success. The main challenge is to know how to design a nonlinear activation function inside the quantum neuron, because the laws in quantum mechanics require the operations on quantum neurons be unitary and linear. A recent discovery uses a special quantum circuit technique called repeat-until-success to make a nonlinear activation function inside a quantum neuron, which is the hard part of creating this neuron. However, the activation function used in that work is based on the periodic tangent function. Because of this periodicity, the input to this function has to be restricted to the range of [0, π/2), which is a serious constraint for its applications in real world problems. The function’s periodicity also makes its neurons not suited for being trained with gradient descent as its derivatives oscillate. The purpose of our study is to propose a new nonlinear activation function that is not periodic so it can take any real numbers and its neurons can be trained with efficient gradient descent. Our quantum neuron offers the full benefit as a quantum entity to support superposition, entanglement, interference, while also enjoys the full benefit as a classical entity to take any real numbers as its input and can be trained with gradient descent. The performance of the quantum neurons with our new activation function is analyzed on IBM’s 5Q quantum computer and IBM’s quantum simulator.

Natural Science•Jan 1, 2018•27 citations
All-optical neural networks (AONNs) promise transformative gains in speed and energy efficiency for artificial intelligence (AI) by leveraging light's intrinsic parallelism and wave nature. However, their scalability has been fundamentally limited by the high power requirements of conventional nonlinear optical elements. Here, we present a low-power nonlinear activation scheme based on a three-level quantum system driven by dual laser fields. This platform introduces a two-channel nonlinear activation matrix with self- and cross-nonlinearities, enabling true multi-input, multi-output optical processing. The system supports tunable activation behaviors, including sigmoid and ReLU functions, at ultralow power levels (17 μ W per neuron). We validate our approach through theoretical modeling and experimental demonstration in rubidium vapor cells, showing the feasibility of scaling to deep AONNs with millions of neurons operating under 20 W of total optical power. Crucially, we demonstrate the all-optical generation of gradient-like signals with backpropagation, paving the way for all-optical training. These results mark a significant advance toward scalable, high-speed, and energy-efficient optical AI hardware.
Optical neural networks promise ultrafast, low-energy information processing by performing computation directly with photons. Current implementations, however, are largely restricted to steady-state operation and rely on high-latency electro-optical conversion for nonlinear activation. To address these limitations, we propose an all-optical fully connected neural network architecture in which the basic neuronal functions are realized by coherent transient quantum dynamics. Within this framework, phase-tunable nonlocal interference in a giant cavity implements programmable synaptic weights; an integrator operating in the bad cavity regime performs temporal summation by coherently combining sequential wavepackets; and transient Rabi dynamics of a driven two-level system provide nonlinear activation. Full-physics simulations demonstrate high classification accuracy on MNIST and colored-object recognition tasks. These results eliminate the optoelectronic activation bottleneck, reduce latency, and establish transient light-matter dynamics as a native physical resource for high-dimensional nonlinear information processing, paving the way toward fully optical neuromorphic computing.
Optical neural networks (ONNs) have been developed to enhance processing speed and energy efficiency in machine learning by leveraging optical devices for nonlinear activation and establishing connections among neurons. In this work, we propose a quantum optical neural network (QONN) that utilizes atom-cavity neurons with controllable photon absorption and emission. These quantum neurons are designed to replace the electronic components in ONNs, which typically introduce delays and substantial energy consumption during nonlinear activation. To evaluate the performance of the QONN, we apply it to the MNIST digit classification task, considering the effects of photon absorption duration, random atom-cavity detuning, and stochastic photon loss. Additionally, we introduce a convolutional QONN to facilitate a real-world satellite image classification (SAT-6) task. Due to its compact hardware and low power consumption, the QONN offers a promising solution for real-time satellite sensing, reducing communication bandwidth with ground stations and thereby enhancing data security.
Encoding quantum information within bosonic modes offers a promising direction for hardware-efficient and fault-tolerant quantum information processing. However, achieving high-fidelity universal control over bosonic encodings using native photonic hardware remains a significant challenge. We establish a quantum control framework to prepare and perform universal logical operations on arbitrary multimode multi-photon states using a quantum photonic neural network. Central to our approach is the optical nonlinearity, which is realized through strong light-matter interaction with a three-level Λ atomic system. The dynamics of this passive interaction are asymptotically confined to the single-mode subspace, enabling the construction of deterministic entangling gates and overcoming limitations faced by many nonlinear optical mechanisms. Using this nonlinearity as the element-wise activation function, we show that the proposed architecture is able to deterministically prepare a wide array of multimode multi-photon states, including essential resource states. We demonstrate universal code-agnostic control of bosonic encodings by preparing and performing logical operations on symmetry-protected error-correcting codes. Our architecture is not constrained by symmetries imposed by evolution under a system Hamiltonian such as purely χ (2) and χ (3) processes, and is naturally suited to implement non-transversal gates on photonic logical qubits. Additionally, we propose an error-correction scheme based on non-demolition measurements that is facilitated by the optical nonlinearity as a building block. Our results pave the way for near-term quantum photonic processors that enable error-corrected quantum computation, and can be achieved using present-day integrated photonic hardware.
Continuous-variables (CV) quantum optics is a natural formalism for neural networks (NNs) due to its ability to reproduce the information processing of such trainable interconnected systems. In quantum optics, Gaussian operators induce affine mappings on the quadratures of optical modes while non-Gaussian resources (the challenging piece for physical implementation) originate the nonlinear effects, unlocking quantum analogs of an artificial neuron. This work presents a novel experimentally-feasible framework for continuous-variable quantum optical neural networks (QONNs) developed with available photonic components: coherent states as input encoding, a general Gaussian transformation followed by multi-mode photon subtractions as the processing layer, and homodyne detection as outputs readout. The closed-form expressions of such architecture are derived demonstrating the family of adaptive activations and the quantum-optical neurons that emerge from the amount of photon-subtracted modes, proving that the proposed design satisfies the Universal Approximation Theorem within a single layer. To classically simulate the QONN training, the high-performance QuaNNTO library has been developed based on Wick-Isserlis expansion and Bogoliubov transformations, allowing multi-layer exact expectation values of non-Gaussian states without truncating the infinite-dimensional Hilbert space. Experiments on supervised learning and state-preparation tasks show balanced-resource efficiency with strong expressivity and generalization capabilities, illustrating the potential of the architecture for scalable photonic quantum machine learning and for quantum applications such as complex non-Gaussian gate synthesis.
Two strong transformative fields in contemporary artificial intelligence (AI) are quantum computing and deep learning. Neuro-nanophotonic systems also introduce quantum-inspired innovations that enhance the computing power of neural networks. Quantum-Inspired Neural Networks (QINNs) have shown promise as a potential alternative to the classical deep learning models, particularly when processing high-dimensional data and solving complex problems efficiently. QINNs leverage the attributes of quantum mechanics to enable efficient deep learning methodologies, achieving superior performance in AI-based applications. In this research, we present a QINN framework that synergizes the power of Quantum Variational Circuits (QVCs) and Quantum Natural Gradient Descent (QNGD) to improve learning efficiency of learning. Experimental results show that QINNs can achieve 97% accuracy, which is 4.5% improvement over the classical models, and computational time and energy consumption can be reduced by 45% and 53%, respectively. Furthermore, QINNs exhibited greater robustness to adversarial noise, making them better candidates for healthcare, cybersecurity, and large-scale AI applications. The results illuminate the game-changing ability of QINNs with respect to scalable, cross-computational and highly resilient AI architectures.
Google’s AlphaGo represents the impressive performance of deep learning and the backbone of deep learning is the workhorse of highly versatile neural networks. Each network is made up of layers of interconnected neurons and the nonlinear activation function inside each neuron is one of the key factors that account for the unprecedented achievement of deep learning. Learning how to create quantum neural networks has been a long time pursuit since 1990’s from many researchers, unfortunately without much success. The main challenge is to know how to design a nonlinear activation function inside the quantum neuron, because the laws in quantum mechanics require the operations on quantum neurons be unitary and linear. A recent discovery uses a special quantum circuit technique called repeat-until-success to make a nonlinear activation function inside a quantum neuron, which is the hard part of creating this neuron. However, the activation function used in that work is based on the periodic tangent function. Because of this periodicity, the input to this function has to be restricted to the range of [0, π/2), which is a serious constraint for its applications in real world problems. The function’s periodicity also makes its neurons not suited for being trained with gradient descent as its derivatives oscillate. The purpose of our study is to propose a new nonlinear activation function that is not periodic so it can take any real numbers and its neurons can be trained with efficient gradient descent. Our quantum neuron offers the full benefit as a quantum entity to support superposition, entanglement, interference, while also enjoys the full benefit as a classical entity to take any real numbers as its input and can be trained with gradient descent. The performance of the quantum neurons with our new activation function is analyzed on IBM’s 5Q quantum computer and IBM’s quantum simulator.