In recent years, pre-trained transformer-based language models (LM) have become a key resource for implementing most NLP tasks. However, pre-training such models demands large text collections not available in most languages. In this paper, we study the use of machine-translated corpora for pre-training LMs. We answer the following research questions: RQ1: Is MT-based data an alternative to real data for learning a LM?; RQ2: Can real data be complemented with translated data and improve the resulting LM? In order to validate these two questions, several BERT models for Basque have been trained, combining real data and synthetic data translated from Spanish.The evaluation carried out on 9 NLU tasks indicates that models trained exclusively on translated data offer competitive results. Furthermore, models trained with real data can be improved with synthetic data, although further research is needed on the matter.

Large language models are very resource intensive, both financially and environmentally, and require an amount of training data which is simply unobtainable for the majority of NLP practitioners. Previous work has researched the scaling laws of such models, but optimal ratios of model parameters, dataset size, and computation costs focused on the large scale. In contrast, we analyze the effect those variables have on the performance of language models in constrained settings, by building three lightweight BERT models (16M/51M/124M parameters) trained over a set of small corpora (5M/25M/125M words).We experiment on four languages of different linguistic characteristics (Basque, Spanish, Swahili and Finnish), and evaluate the models on MLM and several NLU tasks. We conclude that the power laws for parameters, data and compute for low-resource settings differ from the optimal scaling laws previously inferred, and data requirements should be higher. Our insights are consistent across all the languages we study, as well as across the MLM and downstream tasks. Furthermore, we experimentally establish when the cost of using a Transformer-based approach is worth taking, instead of favouring other computationally lighter solutions.

Natural Language Understanding (NLU) technology has improved significantly over the last few years and multitask benchmarks such as GLUE are key to evaluate this improvement in a robust and general way. These benchmarks take into account a wide and diverse set of NLU tasks that require some form of language understanding, beyond the detection of superficial, textual clues. However, they are costly to develop and language-dependent, and therefore they are only available for a small number of languages. In this paper, we present BasqueGLUE, the first NLU benchmark for Basque, a less-resourced language, which has been elaborated from previously existing datasets and following similar criteria to those used for the construction of GLUE and SuperGLUE. We also report the evaluation of two state-of-the-art language models for Basque on BasqueGLUE, thus providing a strong baseline to compare upon. BasqueGLUE is freely available under an open license.

Until recently, coreference resolution has been a critical task on the pipeline of any NLP task involving deep language understanding, such as machine translation, chatbots, summarization or sentiment analysis. However, nowadays, those end tasks are learned end-to-end by deep neural networks without adding any explicit knowledge about coreference. Thus, coreference resolution is used less in the training of other NLP tasks or trending pretrained language models. In this paper we present a new approach to face coreference resolution as a sequence to sequence task based on the Transformer architecture. This approach is simple and universal, compatible with any language or dataset (regardless of singletons) and easier to integrate with current language models architectures. We test it on the ARRAU corpus, where we get 65.6 F1 CoNLL. We see this approach not as a final goal, but a means to pretrain sequence to sequence language models (T5) on coreference resolution.

In this paper, we present a cross-lingual neural coreference resolution system for a less-resourced language such as Basque. To begin with, we build the first neural coreference resolution system for Basque, training it with the relatively small EPEC-KORREF corpus (45,000 words). Next, a cross-lingual coreference resolution system is designed. With this approach, the system learns from a bigger English corpus, using cross-lingual embeddings, to perform the coreference resolution for Basque. The cross-lingual system obtains slightly better results (40.93 F1 CoNLL) than the monolingual system (39.12 F1 CoNLL), without using any Basque language corpus to train it.