Pretrained large character-level language models have been recently
revi...
Subword-level models have been the dominant paradigm in NLP. However,
ch...
Character-based representations have important advantages over subword-b...
Massively multilingual models are promising for transfer learning across...
This paper investigates very low resource language model pretraining, wh...
This paper describes the methods behind the systems submitted by the
Uni...
Recent advances in the field of multilingual dependency parsing have bro...
The transformer-based pre-trained language model BERT has helped to impr...
We propose MoNoise: a normalization model focused on generalizability an...
We present an approach to learning multi-sense word embeddings relying b...
Word representations induced from models with discrete latent variables
...