latent spaces

Latent spaces are abstract representations of data learned by models, where similar items are positioned closer together. In generative models like VAEs or GANs, these spaces help capture the essence of the data, enabling tasks like interpolation and generation of new samples.

8 papers