Quantum Down Sampling Filter for Variational Auto-encoder
This work addresses image reconstruction quality for generative modeling, but it appears incremental as it builds on existing VAE frameworks with quantum enhancements.
The paper tackles the problem of low fidelity in variational autoencoder reconstructions by introducing a quantum variational autoencoder (Q-VAE) that integrates quantum encoding, and it shows that Q-VAE outperforms classical VAEs on MNIST and USPS datasets with lower Fréchet inception distance scores.
Variational autoencoders (VAEs) are fundamental for generative modeling and image reconstruction, yet their performance often struggles to maintain high fidelity in reconstructions. This study introduces a hybrid model, quantum variational autoencoder (Q-VAE), which integrates quantum encoding within the encoder while utilizing fully connected layers to extract meaningful representations. The decoder uses transposed convolution layers for up-sampling. The Q-VAE is evaluated against the classical VAE and the classical direct-passing VAE, which utilizes windowed pooling filters. Results on the MNIST and USPS datasets demonstrate that Q-VAE consistently outperforms classical approaches, achieving lower Fréchet inception distance scores, thereby indicating superior image fidelity and enhanced reconstruction quality. These findings highlight the potential of Q-VAE for high-quality synthetic data generation and improved image reconstruction in generative models.