Xabier Saralegi's research while affiliated with Elhuyar Fundazioa and other places

Context-aware Neural Machine Translation can potentially enhance automated translation quality through effective modelling of context beyond the sentence level. However, suitable corpora for contextual modelling are still scarce, presenting a significant challenge for the training and evaluation of context-aware systems. To address this challenge, we describe \textsc{tando^+}, a document-level corpus for the under-resourced language pairs Basque-French and Basque-Spanish. We provide a detailed description of this corpus, which is to be shared with the scientific community. The corpus comprises parallel data from diverse domains (literature, subtitles, and news) and incorporates context-level information. Additionally, it provides manually crafted contrastive test sets for Basque-Spanish, designed for comprehensive assessment of gender and register contextual phenomena. Additionally, we train and evaluate sentence-level baseline models and several state-of-the-art contextual variants. Our results and analyses indicate that the corpus is well-suited to train and evaluate context-aware machine translation systems for the two selected under-resourced language pairs.

Biosanitary experts around the world are directing their efforts towards the study of COVID-19. This effort generates a large volume of scientific publications at a speed that makes the effective acquisition of new knowledge difficult. Therefore, Information Systems are needed to assist biosanitary experts in accessing, consulting and analyzing these publications. In this work we develop a study of the variables involved in the development of a Question Answering system that receives a set of questions asked by experts about the disease COVID-19 and its causal virus SARS-CoV-2, and provides a ranked list of expert-level answers to each question. In particular, we address the interrelation of the Information Retrieval and the Answer Extraction steps. We found that a recall based document retrieval that leaves to a neural answer extraction module the scanning of the whole documents to find the best answer is a better strategy than relying in a precise passage retrieval before extracting the answer span.

This work analyzes the feasibility of training a neural retrieval system for a collection of scientific papers about COVID-19 using pseudo-qrels extracted from the collection. We propose a method for generating pseudo-qrels that exploits two characteristics present in scientific articles: a) the relationship between title and abstract, and b) the relationship between articles through sentences containing citations. Through these signals we generate pseudo-queries and their respective pseudo-positive (relevant documents) and pseudo-negative (non-relevant documents) examples. The article retrieval process combines a ranking model based on term-maching techniques and a neural one based on pretrained BERT models. BERT models are fine-tuned to the task using the pseudo-qrels generated. We compare different BERT models, both open domain and biomedical domain, and also the generated pseudo-qrels with the open domain MS-Marco dataset for fine-tuning the models. The results obtained on the TREC-COVID collection show that pseudo-qrels provide a significant improvement to neural models, both against classic IR baselines based on term-matching and neural systems trained on MS-Marco.

Word embeddings and pre-trained language models allow to build rich representations of text and have enabled improvements across most NLP tasks. Unfortunately they are very expensive to train, and many small companies and research groups tend to use models that have been pre-trained and made available by third parties, rather than building their own. This is suboptimal as, for many languages, the models have been trained on smaller (or lower quality) corpora. In addition, monolingual pre-trained models for non-English languages are not always available. At best, models for those languages are included in multilingual versions, where each language shares the quota of substrings and parameters with the rest of the languages. This is particularly true for smaller languages such as Basque. In this paper we show that a number of monolingual models (FastText word embeddings, FLAIR and BERT language models) trained with larger Basque corpora produce much better results than publicly available versions in downstream NLP tasks, including topic classification, sentiment classification, PoS tagging and NER. This work sets a new state-of-the-art in those tasks for Basque. All benchmarks and models used in this work are publicly available.

Talaia is a platform for monitoring social media and digital press. A configurable crawler gathers content with respect to user defined domains or topics. Crawled data is processed by means of IXA-pipes NLP chain and EliXa sentiment analysis system. A Django powered interface provides data visualization to provide the user analysis of the data. This paper presents the architecture of the system and describes in detail the different components of the system. To prove the validity of the approach, two real use cases are accounted for, one in the cultural domain and one in the political domain. Evaluation for the sentiment analysis task in both scenarios is also provided, showing the capacity for domain adaptation.

This paper presents a supervised Aspect Based Sentiment Analysis (ABSA) system. Our aim is to develop a modular platform which allows to easily conduct experiments by replacing the modules or adding new features. We obtain the best result in the Opinion Target Extraction (OTE) task (slot 2) using an off-the-shelf sequence labeler. The target polarity classification (slot 3) is addressed by means of a multiclass SVM algorithm which includes lexical based features such as the polarity values obtained from domain and open polarity lexicons. The system obtains accuracies of 0.70 and 0.73 for the restaurant and laptop domain respectively, and performs second best in the out-of-domain hotel, achieving an accuracy of 0.80.

This paper presents a supervised Aspect Based Sentiment Analysis (ABSA) system. Our aim is to develop a modular platform which allows to easily conduct experiments by replacing the modules or adding new features. We obtain the best result in the Opinion Target Extraction (OTE) task (slot 2) using an off-the-shelf sequence labeler. The target polarity classification (slot 3) is addressed by means of a multiclass SVM algorithm which includes lexical based features such as the polarity values obtained from domain and open polarity lexicons. The system obtains accuracies of 0.70 and 0.73 for the restaurant and laptop domain respectively, and performs second best in the out-of-domain hotel, achieving an accuracy of 0.80.