<ul><li>Unsupervised representation learning is essential for applications like Neural Architecture Search (NAS).</li><li>Variational Autoencoders (VAEs) often result in a high percentage of invalid or duplicate architectures when sampling from the continuous representation space.</li><li>A Vector Quantized Variational Autoencoder (VQ-VAE) is introduced to learn a discrete latent space for neural architectures.</li><li>The VQ-VAE approach significantly improves the generation of valid and unique architectures in NAS.</li></ul>

Arch-LLM: Taming LLMs for Neural Architecture Generation via Unsupervised Discrete Representation Learning

Discover more