For machine translation to tackle discourse phenomena, models must have access to extrasentential linguistic context. There has been recent interest in modelling context in neural machine translation (NMT), but models have been principally evaluated with standard automatic metrics, poorly adapted to evaluating discourse phenomena. In this article, we present hand-crafted, discourse test sets, designed to test the models' ability to exploit previous source and target sentences. We investigate the performance of recently proposed multi-encoder NMT models trained on subtitles for English to French. We also explore a novel way of exploiting context from the previous sentence. Despite gains using BLEU, multi-encoder models give limited improvement in the handling of discourse phenomena: 50% accuracy on our coreference test set and 53.5% for coherence/cohesion (compared to a non-contextual baseline of 50%). A simple strategy of decoding the concatenation of the previous and current sentence leads to good performance, and our novel strategy of multiencoding and decoding of two sentences leads to the best performance (72.5% for coreference and 57% for coherence/cohesion), highlighting the importance of target-side context.

Large language models have recently been shown to attain reasonable zero-shot generalization on a diverse set of tasks (Brown et al., 2020). It has been hypothesized that this is a consequence of implicit multitask learning in language model training . Can zero-shot generalization instead be directly induced by explicit multitask learning? To test this question at scale, we develop a system for easily mapping general natural language tasks into a human-readable prompted form. We convert a large set of supervised datasets, each with multiple prompts using varying natural language. These prompted datasets allow for benchmarking the ability of a model to perform completely unseen tasks specified in natural language. We fine-tune a pretrained encoder-decoder model on this multitask mixture covering a wide variety of tasks. The model attains strong zero-shot performance on several standard datasets, often outperforming models up to 16× its size. Further, our approach attains strong performance on a subset of tasks from the BIG-Bench benchmark, outperforming models up to 6× its size. All prompts and trained models are available at github.com/bigscience-workshop/promptsource/ and huggingface.co/bigscience/T0pp.

No abstract

The University of Edinburgh participated in the WMT19 Shared Task on News Translation in six language directions: English↔Gujarati, English↔Chinese, German→English, and English→Czech. For all translation directions, we created or used back-translations of monolingual data in the target language as additional synthetic training data. For English↔Gujarati, we also explored semisupervised MT with cross-lingual language model pre-training, and translation pivoting through Hindi. For translation to and from Chinese, we investigated character-based tokenisation vs. sub-word segmentation of Chinese text. For German→English, we studied the impact of vast amounts of back-translated training data on translation quality, gaining a few additional insights over Edunov et al. (2018). For English→Czech, we compared different pre-processing and tokenisation regimes.

The author list is alphabetical and does not reflect the respective author contributions. The task was coordinated by Mariana Neves. sions were judged to be better than the reference translations, for instance, for de/en, en/es and es/en. 10 https://github.com/glample/fastBPE 11 1 encoder layer, 1 decoder layer, both with with GRU cells, embedding dimension of 512, hidden state of dimension 1024, using layer normalization, implemented using Marian NMT and trained using the Adam optimizer.