Document information extraction tasks performed by humans create data consisting of a PDF or document image input, and extracted string outputs.
We present an autoencoder that leverages learned representations to better measure similarities in data space.
Deep generative models trained with large amounts of unlabelled data have proven to be powerful within the domain of unsupervised learning.
Deep generative models parameterized by neural networks have achieved state-of-the-art performance in unsupervised and semi-supervised learning.
We present hash embeddings, an efficient method for representing words in a continuous vector form.
Humans possess an ability to abstractly reason about objects and their interactions, an ability not shared with state-of-the-art deep learning models
This paper takes a step towards temporal reasoning in a dynamically changing video in a latent space.
