With the rise of AI-based text generators, the need for effective detection mechanisms has become paramount. This paper presents new techniques for building adaptable models and optimizing training aspects for identifying synthetically produced texts across multiple generators and domains. The study, divided into binary and multilabel classification tasks, avoids overfitting through strategic training data limitation. A key innovation is the incorporation of multimodal models that blend numerical text features with conventional NLP approaches. The work also delves into optimizing ensemble model combinations via various voting methods, focusing on accuracy as the official metric. The optimized ensemble strategy demonstrates significant efficacy in both subtasks, highlighting the potential of multimodal and ensemble methods in enhancing the robustness of detection systems against emerging text generators.

Nowadays, intimacy is a fundamental aspect of how we relate to other people in social settings. The most frequent way in which we can determine a high level of intimacy is in the use of certain emoticons, curse words, verbs, etc. This paper presents the approach developed to solve SemEval 2023 task 9: Multiligual Tweet Intimacy Analysis. To address the task, a transfer learning approach was conducted by fine tuning various pre-trained languagemodels. Since the dataset supplied by the organizer was highly imbalanced, our main strategy to obtain high prediction values was the implementation of different oversampling and undersampling techniques on the training set. Our final submission achieved an overall Pearson’s r of 0.497.

This work details our approach for addressing Tasks A and B of the Semeval 2023 Task 10: Explainable Detection of Online Sexism (EDOS). For Task A a simple ensemble based of majority vote system was presented. To build our proposal, first a review of transformers was carried out and the 3 best performing models were selected to be part of the ensemble. Next, for these models, the best hyperpameters were searched using a reduced data set. Finally, we trained these models using more data. During the development phase, our ensemble system achieved an f1-score of 0.8403. For task B, we developed a model based on the deBERTa transformer, utilizing the hyperparameters identified for task A. During the development phase, our proposed model attained an f1-score of 0.6467. Overall, our methodology demonstrates an effective approach to the tasks, leveraging advanced machine learning techniques and hyperparameters searches to achieve high performance in detecting and classifying instances of sexism in online text.

This paper presents the approaches proposedfor I2C Group to address the SemEval-2023Task 4: Identification of Human Values behindArguments (ValueEval)”, whose goal is to classify 20 different categories of human valuesgiven a textual argument. The dataset of thistask consists of one argument per line, including its unique argument ID, conclusion, stanceof the premise towards the conclusion and thepremise text. To indicate whether the argumentdraws or not on that category a binary indication (1 or 0) is included. Participants can submit approaches that detect one, multiple, or allof these values in arguments. The task providesan opportunity for researchers to explore theuse of automated techniques to identify humanvalues in text and has potential applications invarious domains such as social science, politics,and marketing. To deal with the imbalancedclass distribution given, our approach undersamples the data. Additionally, the three components of the argument (conclusion, stanceand premise) are used for training. The systemoutperformed the BERT baseline according toofficial evaluation metrics, achieving a f1 scoreof 0.46.

In this paper we present our approach and system description on Task 7a in ProfNer-ST: Identification of profession & occupation in Health related Social Media. Our main contribution is to show the effectiveness of using BETO-Spanish BERT as a model based on transformers pretrained with a Spanish Corpus for classification tasks. In our experiments we compared several architectures based on transformers with others based on classical machine learning algorithms. With this approach, we achieved an F1-score of 0.92 in the evaluation process.

This paper describes the systems developed for I2C Group to participate on Subtasks A and B in English, and Subtask A in Turkish and Arabic in OffensEval (Task 12 of SemEval 2020). In our experiments we compare three architectures we have developed, two based on Transformer and the other based on classical machine learning algorithms. In this paper, the proposed architectures are described, and the results obtained by our systems are presented.

In this paper we present on-going work on annotating negation in Spanish clinical documents. A corpus of anamnesis and radiology reports has been annotated by two domain expert annotators with negation markers and negated events. The Dice coefficient for inter-annotator agreement is higher than 0.94 for negation markers and higher than 0.72 for negated events. The corpus will be publicly released when the annotation process is finished, constituting the first corpus annotated with negation for Spanish clinical reports available for the NLP community.