As pretrained language model emerge and consistently develop, prompt-based training has become a well-studied paradigm to improve the exploitation of models for many natural language processing tasks. Furthermore, prompting demonstrates great performance compared to conventional fine-tuning in scenarios with limited annotated data, such as zero-shot or few-shot situations. Verbalizers are crucial in this context, as they help interpret masked word distributions generated by language models into output predictions. This study introduces a benchmarking approach to assess three common baselines of verbalizers for topic classification in few-shot learning scenarios. Additionally, we find that increasing the number of label words for automatic label word searching enhances model performance. Moreover, we investigate the effectiveness of template assembling with various aggregation strategies to develop stronger classifiers that outperform models trained with individual templates. Our approach achieves comparable results to prior research while using significantly fewer resources. Our code is available at https://github.com/quang-anh-nguyen/verbalizer_benchmark.git.

Information extraction (IE) from text documents is an important NLP task that includes entity, relation, and event extraction. These tasks are often addressed jointly as a graph generation problem, where entities and event triggers represent nodes and where relations and event arguments represent edges. Most existing systems use local classifiers for nodes and edges, trained using cross-entropy loss, and employ inference strategies such as beam search to approximate the optimal graph structure. These approaches typically suffer from exposure bias due to the discrepancy between training and decoding. In this paper, we tackle this problem by casting graph generation as auto-regressive sequence labeling and making its training aware of the decoding procedure by using a differentiable version of beam search. We evaluate the effectiveness of our approach through extensive experiments conducted on the ACE05 and ConLL04 datasets across diverse languages. Our experimental findings affirm that our model outperforms its non-decoding-aware version for all datasets employed. Furthermore, we conduct ablation studies that emphasize the effectiveness of aligning training and inference. Additionally, we introduce a novel quantification of exposure bias within this context, providing valuable insights into the functioning of our model.

Named Entity Recognition (NER) is essential in various Natural Language Processing (NLP) applications. Traditional NER models are effective but limited to a set of predefined entity types. In contrast, Large Language Models (LLMs) can extract arbitrary entities through natural language instructions, offering greater flexibility. However, their size and cost, particularly for those accessed via APIs like ChatGPT, make them impractical in resource-limited scenarios. In this paper, we introduce a compact NER model trained to identify any type of entity. Leveraging a bidirectional transformer encoder, our model, GLiNER, facilitates parallel entity extraction, an advantage over the slow sequential token generation of LLMs. Through comprehensive testing, GLiNER demonstrate strong performance, outperforming both ChatGPT and fine-tuned LLMs in zero-shot evaluations on various NER benchmarks.

Semi-Markov CRF has been proposed as an alternative to the traditional Linear Chain CRF for text segmentation tasks such as Named Entity Recognition (NER). Unlike CRF, which treats text segmentation as token-level prediction, Semi-CRF considers segments as the basic unit, making it more expressive. However, Semi-CRF suffers from two major drawbacks: (1) quadratic complexity over sequence length, as it operates on every span of the input sequence, and (2) inferior performance compared to CRF for sequence labeling tasks like NER. In this paper, we introduce Filtered Semi-Markov CRF, a variant of Semi-CRF that addresses these issues by incorporating a filtering step to eliminate irrelevant segments, reducing complexity and search space. Our approach is evaluated on several NER benchmarks, where it outperforms both CRF and Semi-CRF while being significantly faster. The implementation of our method is available on Github.

In this paper, we study the transferability of Named Entity Recognition (NER) models between Arabic dialects. This question is important because the available manually-annotated resources are not distributed equally across dialects: Modern Standard Arabic (MSA) is much richer than other dialects for which little to no datasets exist. How well does a NER model, trained on MSA, perform on other dialects? To answer this question, we construct four datasets. The first is an MSA dataset extracted from the ACE 2005 corpus. The others are datasets for Egyptian, Morocan and Syrian which we manually annotate following the ACE guidelines. We train a span-based NER model on top of a pretrained language model (PLM) encoder on the MSA data and study its performance on the other datasets in zero-shot settings. We study the performance of multiple PLM encoders from the literature and show that they achieve acceptable performance with no annotation effort. Our annotations and models are publicly available (https://github.com/niamaelkhbir/Arabic-Cross-Dialectal-NER).

The Wojood Named Entity Recognition (NER) shared task introduces a comprehensive Arabic NER dataset encompassing both flat and nested entity tasks, addressing the challenge of limited Arabic resources. In this paper, we present our team LIPN approach to addressing the two subtasks of WojoodNER SharedTask. We frame NER as a span classification problem. We employ a pretrained language model for token representations and neural network classifiers. We use global decoding for flat NER and a greedy strategy for nested NER. Our model secured the first position in flat NER and the fourth position in nested NER during the competition, with an F-score of 91.96 and 92.45 respectively. Our code is publicly available (https://github.com/niamaelkhbir/LIPN-at-WojoodSharedTask).

Nous présentons une extension aux analyseurs syntaxiques en constituants neuronaux modernes qui consiste à doter les constituants potentiels d’une représentation vectorielle affinée en fonction du contexte par plusieurs applications successives d’un module de type transformer efficace (pooling par attention puis transformation non-linéaire).Nous appliquons cette extension à l’analyseur CRF de Yu Zhang & Al.Expérimentalement, nous testons cette extension sur deux corpus (PTB et FTB) avec ou sans vecteurs de mots dynamiques: cette extension permet d’avoir un gain constant dans toutes les configurations.

La reconnaissance d’entités nommées est une tâche importante en traitement automatique du langage naturel avec des applications dans de nombreux domaines. Dans cet article, nous décrivons une nouvelle approche pour la reconnaissance d’entités nommées, dans laquelle nous produisons un ensemble de segmentations en maximisant un score global. Pendant l’entraînement, nous optimisons notre modèle en maximisant la probabilité de la segmentation correcte. Pendant l’inférence, nous utilisons la programmation dynamique pour sélectionner la meilleure segmentation avec une complexité linéaire. Nous prouvons que notre approche est supérieure aux modèles champs de Markov conditionnels et semi-CMC pour la reconnaissance d’entités nommées.

There are two main paradigms for Named Entity Recognition (NER): sequence labelling and span classification. Sequence labelling aims to assign a label to each word in an input text using, for example, BIO (Begin, Inside and Outside) tagging, while span classification involves enumerating all possible spans in a text and classifying them into their labels. In contrast to sequence labelling, unconstrained span-based methods tend to assign entity labels to overlapping spans, which is generally undesirable, especially for NER tasks without nested entities. Accordingly, we propose GNNer, a framework that uses Graph Neural Networks to enrich the span representation to reduce the number of overlapping spans during prediction. Our approach reduces the number of overlapping spans compared to strong baseline while maintaining competitive metric performance. Code is available at https://github.com/urchade/GNNer.

Named Entity Recognition (NER) is an important task in Natural Language Processing with applications in many domains. While the dominant paradigm of NER is sequence labelling, span-based approaches have become very popular in recent times but are less well understood. In this work, we study different aspects of span-based NER, namely the span representation, learning strategy, and decoding algorithms to avoid span overlap. We also propose an exact algorithm that efficiently finds the set of non-overlapping spans that maximizes a global score, given a list of candidate spans. We performed our study on three benchmark NER datasets from different domains. We make our code publicly available at https://github.com/urchade/span-structured-prediction.

Named Entity Recognition (NER) is an important task in Natural Language Processing with applications in many domains. In this paper, we describe a novel approach to named entity recognition, in which we output a set of spans (i.e., segmentations) by maximizing a global score. During training, we optimize our model by maximizing the probability of the gold segmentation. During inference, we use dynamic programming to select the best segmentation under a linear time complexity. We prove that our approach outperforms CRF and semi-CRF models for Named Entity Recognition. We will make our code publicly available.

Like most natural language understanding and generation tasks, state-of-the-art models for summarization are transformer-based sequence-to-sequence architectures that are pretrained on large corpora. While most existing models focus on English, Arabic remains understudied. In this paper we propose AraBART, the first Arabic model in which the encoder and the decoder are pretrained end-to-end, based on BART. We show that AraBART achieves the best performance on multiple abstractive summarization datasets, outperforming strong baselines including a pretrained Arabic BERT-based model, multilingual BART, Arabic T5, and a multilingual T5 model. AraBART is publicly available.

Previous work on Arabic information extraction has mainly focused on named entity recognition and very little work has been done on Arabic relation extraction and event recognition. Moreover, modeling Arabic data for such tasks is not straightforward because of the morphological richness and idiosyncrasies of the Arabic language. We propose in this article the first neural joint information extraction system for the Arabic language.

In this paper we present a parsing model for projective dependency trees which takes advantage of the existence of complementary dependency annotations which is the case in Arabic, with the availability of CATiB and UD treebanks. Our system performs syntactic parsing according to both annotation types jointly as a sequence of arc-creating operations, and partially created trees for one annotation are also available to the other as features for the score function. This method gives error reduction of 9.9% on CATiB and 6.1% on UD compared to a strong baseline, and ablation tests show that the main contribution of this reduction is given by sharing tree representation between tasks, and not simply sharing BiLSTM layers as is often performed in NLP multitask systems.

Dans cet article, nous nous intéressons à l’extraction d’entités médicales de type symptôme dans les textes biomédicaux. Cette tâche est peu explorée dans la littérature et il n’existe pas à notre connaissance de corpus annoté pour entraîner un modèle d’apprentissage. Nous proposons deux approches faiblement supervisées pour extraire ces entités. Une première est fondée sur la fouille de motifs et introduit une nouvelle contrainte de similarité sémantique. La seconde formule la tache comme une tache d’étiquetage de séquences en utilisant les CRF (champs conditionnels aléatoires). Nous décrivons les expérimentations menées qui montrent que les deux approches sont complémentaires en termes d’évaluation quantitative (rappel et précision). Nous montrons en outre que leur combinaison améliore sensiblement les résultats.

En classification de textes, la plupart des méthodes fondées sur des classifieurs statistiques utilisent des mots, ou des combinaisons de mots contigus, comme descripteurs. Si l’on veut prendre en compte plus d’informations le nombre de descripteurs non contigus augmente exponentiellement. Pour pallier à cette croissance, la fouille de motifs séquentiels permet d’extraire, de façon efficace, un nombre réduit de descripteurs qui sont à la fois fréquents et pertinents grâce à l’utilisation de contraintes. Dans ce papier, nous comparons l’utilisation de motifs fréquents sous contraintes et l’utilisation de motifs -libres, comme descripteurs. Nous montrons les avantages et inconvénients de chaque type de motif.

We present annotation guidelines and a web-based annotation framework developed as part of an effort to create a manually annotated Arabic corpus of errors and corrections for various text types. Such a corpus will be invaluable for developing Arabic error correction tools, both for training models and as a gold standard for evaluating error correction algorithms. We summarize the guidelines we created. We also describe issues encountered during the training of the annotators, as well as problems that are specific to the Arabic language that arose during the annotation process. Finally, we present the annotation tool that was developed as part of this project, the annotation pipeline, and the quality of the resulting annotations.

We present a novel translation quality informed procedure for both extraction and scoring of phrase pairs in PBSMT systems. We reformulate the extraction problem in the supervised learning framework. Our goal is twofold. First, We attempt to take the translation quality into account; and second we incorporating arbitrary features in order to circumvent alignment errors. One-Class SVMs and the Mapping Convergence algorithm permit training a single-class classifier to discriminate between useful and useless phrase pairs. Such classifier can be learned from a training corpus that comprises only useful instances. The confidence score, produced by the classifier for each phrase pairs, is employed as a selection criteria. The smoothness of these scores allow a fine control over the size of the resulting translation model. Finally, confidence scores provide a new accuracy-based feature to score phrase pairs. Experimental evaluation of the method shows accurate assessments of phrase pairs quality even for regions in the space of possible phrase pairs that are ignored by other approaches. This enhanced evaluation of phrase pairs leads to improvements in the translation performance as measured by BLEU.

Dans les systèmes de traduction statistique à base de segments, le modèle de traduction est estimé à partir d’alignements mot-à-mot grâce à des heuristiques d’extraction et de valuation. Bien que ces alignements mot-à-mot soient construits par des modèles probabilistes, les processus d’extraction et de valuation utilisent ces modèles en faisant l’hypothèse que ces alignements sont déterministes. Dans cet article, nous proposons de lever cette hypothèse en considérant l’ensemble de la matrice d’alignement, d’une paire de phrases, chaque association étant valuée par sa probabilité. En comparaison avec les travaux antérieurs, nous montrons qu’en utilisant un modèle exponentiel pour estimer de manière discriminante ces probabilités, il est possible d’obtenir des améliorations significatives des performances de traduction. Ces améliorations sont mesurées à l’aide de la métrique BLEU sur la tâche de traduction de l’arabe vers l’anglais de l’évaluation NIST MT’09, en considérant deux types de conditions selon la taille du corpus de données parallèles utilisées.

The quality of statistical machine translation systems depends on the quality of the word alignments that are computed during the translation model training phase. IBM alignment models, as implemented in the GIZA++ toolkit, constitute the de facto standard for performing these computations. The resulting alignments and translation models are however very noisy, and several authors have tried to improve them. In this work, we propose a simple and effective approach, which considers alignment as a series of independent binary classification problems in the alignment matrix. Through extensive feature engineering and the use of stacking techniques, we were able to obtain alignments much closer to manually defined references than those obtained by the IBM models. These alignments also yield better translation models, delivering improved performance in a large scale Arabic to English translation task.