Abstract: Sign language translation (SLT) traditionally requires costly human gloss annotations. Recently, gloss-free approaches, which directly generate text from video, have been studied and ...

Abstract: Machine learning methods offer a shortcut for automated cardiovascular disease diagnosis. However, the high cost of ECG signal annotation, along with insufficient labeled data and class ...

CAESAR is a unified framework for spatio-temporal scientific data reduction. The baseline model, CAESAR-V, is built on a variational autoencoder (VAE) with scale hyperpriors and super-resolution ...

Variational Autoencoders (VAEs) have proven to be powerful generative models in quantitative finance, capable of learning latent representations of market data and generating synthetic scenarios for ...