We introduce MADLAD-400, a manually audited, general domain 3T token
mon...
Reinforcement learning from human feedback (RLHF) can improve the qualit...
Automatic evaluation of machine translation (MT) is a critical tool driv...
State space models (SSMs) have shown impressive results on tasks that re...
The recent rapid progress in pre-training Large Language Models has reli...
Pretrained multilingual large language models have typically used heuris...
Neural machine translation (NMT) has progressed rapidly over the past se...
In this work, we provide a large-scale empirical study of the scaling
pr...
The rapid scaling of language models is motivating research using
low-bi...
We demonstrate the potential of few-shot translation systems, trained wi...
Crosslingual conditional generation (e.g., machine translation) has long...
We present FRMT, a new dataset and evaluation benchmark for Few-shot
Reg...
Recent research has proposed a series of specialized optimization algori...
In this paper we share findings from our effort to build practical machi...
Large language models have been shown to achieve remarkable performance
...
We introduce XTREME-S, a new benchmark to evaluate universal cross-lingu...
Multilingual neural machine translation models are trained to maximize t...
We explore the use of natural language prompts for controlling various
a...
In this work, we study the effect of varying the architecture and traini...
Natural language understanding and generation models follow one of the t...
Achieving universal translation between all human language pairs is the
...
Scaling language models with more data, compute and parameters has drive...
In this work, we study the evolution of the loss Hessian across many
cla...
Document-level neural machine translation (DocNMT) delivers coherent
tra...
This paper explores zero-label learning in Natural Language Processing (...
We present an empirical study of scaling properties of encoder-decoder
T...
Despite the increasing number of large and comprehensive machine transla...
Recent advances in neural machine translation (NMT) have pushed the qual...
In this work, we take the first steps towards building a universal rewri...
Machine learning has brought striking advances in multilingual natural
l...
With the success of large-scale pre-training and multilingual modeling i...
We propose a straightforward vocabulary adaptation scheme to extend the
...
To mitigate the negative effect of low quality training data on the
perf...
One challenge of machine translation is how to quickly adapt to unseen
d...
In this paper, we offer a preliminary investigation into the task of in-...
Multilingual Neural Machine Translation (MNMT) models are commonly train...
Pre-trained cross-lingual encoders such as mBERT (Devlin et al., 2019) a...
Massively multilingual models subsuming tens or even hundreds of languag...
Unsupervised translation has reached impressive performance on resource-...
Neural network scaling has been critical for improving the model quality...
Over the last few years two promising research directions in low-resourc...
Much recent progress in applications of machine learning models to NLP h...
Many sequence-to-sequence generation tasks, including machine translatio...
Most neural networks utilize the same amount of compute for every exampl...
Most neural machine translation systems still translate sentences in
iso...
Neural Machine Translation (NMT) models generally perform translation us...
Fine-tuning pre-trained Neural Machine Translation (NMT) models is the
d...
To train neural machine translation models simultaneously on multiple ta...
Multilingual Neural Machine Translation (NMT) models have yielded large
...
The recently proposed massively multilingual neural machine translation ...