Quantum-secure multiparty deep learning
arXiv:2408.05629 · doi:10.1103/k8wg-qmbh
Abstract
Secure multiparty computation enables the joint evaluation of multivariate functions across distributed users while ensuring the privacy of their local inputs. This field has become increasingly urgent due to the exploding demand for computationally intensive deep learning inference. These computations are typically offloaded to cloud computing servers, leading to vulnerabilities that can compromise the security of the clients' data. To solve this problem, we introduce a linear algebra engine that leverages the quantum nature of light for information-theoretically secure multiparty computation using only conventional telecommunication components. We apply this linear algebra engine to deep learning and derive rigorous upper bounds on the information leakage of both the deep neural network weights and the client's data via the Holevo and the Cramér-Rao bounds, respectively. Applied to the MNIST classification task, we obtain test accuracies exceeding while leaking less than bits per weight symbol and bits per data symbol. This weight leakage is an order of magnitude below the minimum bit precision required for accurate deep learning using state-of-the-art quantization techniques. Our work lays the foundation for practical quantum-secure computation and unlocks secure cloud deep learning as a field.
References in corpus (70)
- Quantum Machine Learning
- The Security of Practical Quantum Key Distribution
- Quantum algorithm for solving linear systems of equations
- Deep Learning with Coherent Nanophotonic Circuits
- All-Optical Machine Learning Using Diffractive Deep Neural Networks
- Quantum support vector machine for big data classification
- Advances in Quantum Cryptography
- Parallel convolution processing using an integrated photonic tensor core
- Photonics for artificial intelligence and neuromorphic computing
- Secure quantum key distribution with realistic devices
- 11 TeraFLOPs per second photonic convolutional accelerator for deep learning optical neural networks
- Quantum Convolutional Neural Networks
- Experimental demonstration of long-distance continuous-variable quantum key distribution
- Neuromorphic Silicon Photonic Networks
- Deep physical neural networks enabled by a backpropagation algorithm for arbitrary physical systems
- Single-chip photonic deep neural network for instantaneous image classification
- Large-scale neuromorphic optoelectronic computing with a reconfigurable diffractive processing unit
- Unconditional optimality of Gaussian attacks against continuous-variable QKD
- Laguerre-Gaussian mode sorter
- Quantum circuits with many photons on a programmable nanophotonic chip
- Experimental Demonstration of Blind Quantum Computing
- Long-Distance Continuous-Variable Quantum Key Distribution over 202.81 km of Fiber
- Unconditional security proof of long-distance continuous-variable quantum key distribution with discrete modulation
- The physics of optical computing
- Security in Quantum Cryptography
- Large-Scale Optical Neural Networks based on Photoelectric Multiplication
- Continuous-Variable Quantum Key Distribution with Gaussian Modulation -- The Theory of Practical Implementations
- Layerwise learning for quantum neural networks
- Experimentally realized in situ backpropagation for deep learning in nanophotonic neural networks
- Extremality of Gaussian quantum states
- Direct measurement of general quantum states using weak measurement
- Unconditionally verifiable blind computation
- An optical neural network using less than 1 photon per multiplication
- Image sensing with multilayer, nonlinear optical neural networks
- Precision metrology using weak measurements
- Delocalized Photonic Deep Learning on the Internet's Edge
- Deep Learning with Coherent VCSEL Neural Networks
- Single chip photonic deep neural network with accelerated training
- Weak measurements are universal
- Experimental quantum speed-up in reinforcement learning agents
- Scalable boson sampling with a single-photon device
- Unconditional security from noisy quantum storage
- Practical continuous-variable quantum key distribution with composable security
- Experimental verification of quantum computations
- Cryptography from Noisy Storage
- Scalable boson-sampling with time-bin encoding using a loop-based architecture
- Asymptotic security analysis of discrete-modulated continuous-variable quantum key distribution
- Covert Optical Communication
- Asymptotic security of continuous-variable quantum key distribution with a discrete modulation
- Universal Blind Quantum Computing with Weak Coherent Pulses
- Experimental Blind Quantum Computing for a Classical Client
- Experimental multiphase estimation on a chip
- Efficient universal blind computation
- Experimental implementation of bit commitment in the noisy-storage model
- An Experimental Implementation of Oblivious Transfer in the Noisy Storage Model
- Experimental plug&play quantum coin flipping
- Finite-Size Security for Discrete-Modulated Continuous-Variable Quantum Key Distribution Protocols
- Quantum-enhanced data classification with a variational entangled sensor network
- Pattern recognition techniques for Boson Sampling validation
- Physical-Layer Supervised Learning Assisted by an Entangled Sensor Network
- Processing entangled photons in high dimensions with a programmable light converter
- All-Fiber Source and Sorter for Multimode Correlated Photons
- Quantum-limited stochastic optical neural networks operating at a few quanta per activation
- High-dimensional coherent one-way quantum key distribution
- Time-Domain Universal Linear-Optical Operations for Universal Quantum Information Processing
- High-dimensional quantum key distribution using a multi-plane light converter
- Composable security in relativistic quantum cryptography
- Secure multi-party quantum computation with few qubits
- Superadditive Communication with the Green Machine: A Practical Demonstration of Nonlocality without Entanglement
- Blind quantum machine learning with quantum bipartite correlator