Abstract: This paper introduces V2Coder, a non-autoregressive vocoder based on hierarchical variational autoencoders (VAEs). The hierarchical VAE with hierarchically extended prior and approximate ...

Abstract: Artificial intelligence (AI) has revolutionized medical imaging, significantly improving diagnostic accuracy via the evaluation of X-rays, MRIs, and CT scans. However, the effectiveness of ...

Builds on Toy Models of Superposition and Towards Monosemanticity: Decomposing Language Models With Dictionary Learning. Addressing concerns that correlated features remain entangled in current SAE ...

This project implements a Variational Autoencoder (VAE) for image generation. Unlike standard autoencoders, VAE learns a probabilistic latent space by encoding images to a distribution and sampling ...