Reproduce “Quantum-Trained Convolutional Neural Network for Deepfake Audio Detection” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Chu-Hsuan Abraham Lin, Chen-Yu Liu, Samuel Yen-Chi Chen, Kuan-Cheng Chen
A Quantum-Trained Convolutional Neural Network (QT-CNN) is introduced for deepfake audio detection, leveraging quantum computation during training to significantly reduce model parameters. This hybrid approach achieved up to 70% parameter reduction while maintaining accuracy levels comparable to classical CNNs.
Key insights
- Quantum-enhanced training can effectively compress neural network models, enabling substantial reductions in the trainable parameters of classical CNNs.
- It is possible to maintain deepfake detection accuracy comparable to classical models while achieving significant parameter reduction through a hybrid quantum-classical approach.
- Quantum advantages can be practically leveraged in the Noisy Intermediate-Scale Quantum (NISQ) era by confining quantum operations to the training phase, thus bypassing current hardware limitations for inference.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.