Recent research has explored methods for updating and modifying factual knowledge in large language models, often focusing on specific multi-layer perceptron blocks. This study expands on this work by examining the effectiveness of existing knowledge editing methods across languages and delving into the role of attention mechanisms in this process. Drawing from the insights gained, we propose Mass-Editing Memory with Attention in Transformers (MEMAT), a method that achieves significant improvements in all metrics while requiring minimal parameter modifications. MEMAT delivers a remarkable 10% increase in magnitude metrics, benefits languages not included in the training data and also demonstrates a high degree of portability. Our code and data are at https://github.com/dtamayo-nlp/MEMAT.

Much work has gone into developing language models of increasing size, but only recently have we begun to examine them for pernicious behaviour that could lead to harming marginalised groups. Following Lin et al. (2022) in rooting our work in psychological research, we prompt two masked language models (MLMs) of different specialisations in English and Spanish with statements from a questionnaire developed to measure stigma to determine if they treat physical and mental illnesses equally. In both models we find a statistically significant difference in the treatment of physical and mental illnesses across most if not all latent constructs as measured by the questionnaire, and thus they are more likely to associate mental illnesses with stigma. We then examine their training data or data retrieved from the same domain using a computational implementation of the Stereotype Content Model (SCM) (Fiske et al., 2002; Fraser et al., 2021) to interpret the questionnaire results based on the SCM values as reflected in the data. We observe that model behaviour can largely be explained by the distribution of the mentions of illnesses according to their SCM values.

We present the findings and results of our pseudonymisation system, which has been developed for a real-life use-case involving users and an informative chatbot in the context of the COVID-19 pandemic. Message exchanges between the two involve the former group providing information about themselves and their residential area, which could easily allow for their re-identification. We create a modular pipeline to detect PIIs and perform basic deidentification such that the data can be stored while mitigating any privacy concerns. The use-case presents several challenging aspects, the most difficult of which is the logistic challenge of not being able to directly view or access the data due to the very privacy issues we aim to resolve. Nevertheless, our system achieves a high recall of 0.99, correctly identifying almost all instances of personal data. However, this comes at the expense of precision, which only reaches 0.64. We describe the sensitive information identification in detail, explaining the design principles behind our decisions. We additionally highlight the particular challenges we’ve encountered.

This paper introduces the methodology of BSC-LANGTECH team for the FIGNEWS 2024 Shared Task on News Media Narratives. Following the bias annotation subtask, we apply the theory and methods of framing analysis to develop guidelines to annotate bias in the corpus provided by the task organizators. The manual annotation of a subset, with which a moderate IAA agreement has been achieved, is further used in Deep Learning techniques to explore automatic annotation and test the reliability of our framework.

We present and describe two language resources in this paper: CATalog 1.0, the largest text corpus in Catalan to date, and CURATE (Corpus Utility for RAting TExt), a modular, parallelizable pipeline used for processing and scoring documents based on text quality that we have optimised to run in High Performance Cluster (HPC) environments. In the coming sections we describe our data preprocessing pipeline at length; traditional pipelines usually implement a set of binary filters such that a given document is either in or out. In our experience with Catalan, in lower-resource settings it is more practical to instead assign a document a soft score to allow for more flexible decision-making. We describe how the document score is calculated and highlight its interpretability by showing that it is significantly correlated with human judgements as obtained from a comparative judgement experiment. We additionally describe the different subcorpora that make up CATalog 1.0.

Current LLM-based applications are becoming steadily available for everyone with a reliable access to technology and the internet. These applications offer benefits to their users that leave those without access to them at a serious disadvantage. Given the vastly large amount of data needed to train LLMs, the gap between languages with access to such quantity of data and those without it is currently larger than ever. Aimed at saving this gap, the Aina Project was created to provide Catalan with the necessary resources to keep being relevant in the context of AI/NLP applications based on LLMs. We thus present a set of strategies to consider when improving technology support for a mid- or low-resource language, specially addressing sustainability of high-quality data acquisition and the challenges involved in the process. We also introduce a large amount of new annotated data for Catalan. Our hope is that those interested in replicating this work for another language can learn from what worked for us, the challenges that we faced, and the sometimes disheartening truth of working with mid- and low-resource languages.

Large language models have amply proven their great capabilities, both in downstream tasks and real-life settings. However, low- and mid-resource languages do not have access to the necessary means to train such models from scratch, and often have to rely on multilingual models despite being underrepresented in the training data. For the particular case of the Catalan language, we prove that continued pre-training with vocabulary adaptation is a better alternative to take the most out of already pre-trained models, even if these have not seen any Catalan data during their pre-training phase. We curate a 26B tokens corpus and use it to further pre-train BLOOM, giving rise to the FLOR models. We perform an extensive evaluation to assess the effectiveness of our method, obtaining consistent gains across Catalan and Spanish tasks. The models, training data, and evaluation framework are made freely available under permissive licenses.

Zero-shot text classification is a widely studied task that deals with a lack of annotated data. The most common approach is to reformulate it as a textual entailment problem, enabling classification into unseen classes. This work explores an effective approach that trains on a weakly supervised dataset generated from traditional classification data. We empirically study the relation between the performance of the entailment task, which is used as a proxy, and the target zero-shot text classification task. Our findings reveal that there is no linear correlation between both tasks, to the extent that it can be detrimental to lengthen the fine-tuning process even when the model is still learning, and propose a straightforward method to stop training on time. As a proof of concept, we introduce a domain-specific zero-shot text classifier that was trained on Microsoft Academic Graph data. The model, called SCIroShot, achieves state-of-the-art performance in the scientific domain and competitive results in other areas. Both the model and evaluation benchmark are publicly available on HuggingFace and GitHub.

This work presents the first large-scale biomedical Spanish language models trained from scratch, using large biomedical corpora consisting of a total of 1.1B tokens and an EHR corpus of 95M tokens. We compared them against general-domain and other domain-specific models for Spanish on three clinical NER tasks. As main results, our models are superior across the NER tasks, rendering them more convenient for clinical NLP applications. Furthermore, our findings indicate that when enough data is available, pre-training from scratch is better than continual pre-training when tested on clinical tasks, raising an exciting research question about which approach is optimal. Our models and fine-tuning scripts are publicly available at HuggingFace and GitHub.

Nested Named Entity Recognition (NER) is an information extraction task that aims to identify entities that may be nested within other entity mentions. Despite the availability of several corpora with nested entities in the Spanish clinical domain, most previous work has overlooked them due to the lack of models and a clear annotation scheme for dealing with the task. To fill this gap, this paper provides an empirical study of straightforward methods for tackling the nested NER task on two Spanish clinical datasets, Clinical Trials, and the Chilean Waiting List. We assess the advantages and limitations of two sequence labeling approaches; one based on Multiple LSTM-CRF architectures and another on Joint labeling models. To better understand the differences between these models, we compute task-specific metrics that adequately measure the ability of models to detect nested entities and perform a fine-grained comparison across models. Our experimental results show that employing domain-specific language models trained from scratch significantly improves the performance obtained with strong domain-specific and general-domain baselines, achieving state-of-the-art results in both datasets. Specifically, we obtained F1 scores of 89.21 and 83.16 in Clinical Trials and the Chilean Waiting List, respectively. Interestingly enough, we observe that the task-specific metrics and analysis properly reflect the limitations of the models when recognizing nested entities. Finally, we perform a case study on an aggregated NER dataset created from several clinical corpora in Spanish. We highlight how entity length and the simultaneous recognition of inner and outer entities are the most critical variables for the nested NER task.

One of the biomedical entity types of relevance for medicine or biosciences are chemical compounds and drugs. The correct detection these entities is critical for other text mining applications building on them, such as adverse drug-reaction detection, medication-related fake news or drug-target extraction. Although a significant effort was made to detect mentions of drugs/chemicals in English texts, so far only very limited attempts were made to recognize them in medical documents in other languages. Taking into account the growing amount of medical publications and clinical records written in Spanish, we have organized the first shared task on detecting drug and chemical entities in Spanish medical documents. Additionally, we included a clinical concept-indexing sub-track asking teams to return SNOMED-CT identifiers related to drugs/chemicals for a collection of documents. For this task, named PharmaCoNER, we generated annotation guidelines together with a corpus of 1,000 manually annotated clinical case studies. A total of 22 teams participated in the sub-track 1, (77 system runs), and 7 teams in the sub-track 2 (19 system runs). Top scoring teams used sophisticated deep learning approaches yielding very competitive results with F-measures above 0.91. These results indicate that there is a real interest in promoting biomedical text mining efforts beyond English. We foresee that the PharmaCoNER annotation guidelines, corpus and participant systems will foster the development of new resources for clinical and biomedical text mining systems of Spanish medical data.

Word embeddings are representations of words in a dense vector space. Although they are not recent phenomena in Natural Language Processing (NLP), they have gained momentum after the recent developments of neural methods and Word2Vec. Regarding their applications in medical and clinical NLP, they are invaluable resources when training in-domain named entity recognition systems, classifiers or taggers, for instance. Thus, the development of tailored word embeddings for medical NLP is of great interest. However, we identified a gap in the literature which we aim to fill in this paper: the availability of embeddings for medical NLP in Spanish, as well as a standardized form of intrinsic evaluation. Since most work has been done for English, some established datasets for intrinsic evaluation are already available. In this paper, we show the steps we employed to adapt such datasets for the first time to Spanish, of particular relevance due to the considerable volume of EHRs in this language, as well as the creation of in-domain medical word embeddings for the Spanish using the state-of-the-art FastText model. We performed intrinsic evaluation with our adapted datasets, as well as extrinsic evaluation with a named entity recognition systems using a baseline embedding of general-domain. Both experiments proved that our embeddings are suitable for use in medical NLP in the Spanish language, and are more accurate than general-domain ones.

This paper describes the upgrading process of the Multilingual Central Repository (MCR). The new MCR uses WordNet 3.0 as Interlingual-Index (ILI). Now, the current version of the MCR integrates in the same EuroWordNet framework wordnets from five different languages: English, Spanish, Catalan, Basque and Galician. In order to provide ontological coherence to all the integrated wordnets, the MCR has also been enriched with a disparate set of ontologies: Base Concepts, Top Ontology, WordNet Domains and Suggested Upper Merged Ontology. The whole content of the MCR is freely available.

WordNet Domains (WND) is a lexical resource where synsets have been semi-automatically annotated with one or more domain labels from a set of 165 hierarchically organized domains. The uses of WND include the power to reduce the polysemy degree of the words, grouping those senses that belong to the same domain. But the semi-automatic method used to develop this resource was far from being perfect. By cross-checking the content of the Multilingual Central Repository (MCR) it is possible to find some errors and inconsistencies. Many are very subtle. Others, however, leave no doubt. Moreover, it is very difficult to quantify the number of errors in the original version of WND. This paper presents a novel semi-automatic method to propagate domain information through the MCR. We also compare both labellings (the original and the new one) allowing us to detect anomalies in the original WND labels.