Natural Language Inference (NLI) evaluation is crucial for assessing language understanding models; however, popular datasets suffer from systematic spurious correlations that artificially inflate actual model performance. To address this, we propose a method for the automated creation of a challenging test set without relying on the manual construction of artificial and unrealistic examples. We categorize the test set of popular NLI datasets into three difficulty levels by leveraging methods that exploit training dynamics. This categorization significantly reduces spurious correlation measures, with examples labeled as having the highest difficulty showing markedly decreased performance and encompassing more realistic and diverse linguistic phenomena. When our characterization method is applied to the training set, models trained with only a fraction of the data achieve comparable performance to those trained on the full dataset, surpassing other dataset characterization techniques. Our research addresses limitations in NLI dataset construction, providing a more authentic evaluation of model performance with implications for diverse NLU applications.

In recent years, Large Language Models (LLMs) have achieved almost human-like performance on various tasks. While some LLMs have been trained on multilingual data, most of the training data is in English; hence, their performance in English greatly exceeds other languages. To our knowledge, we are the first to collect and translate a large collection of texts, instructions, and benchmarks and train, evaluate, and release open-source LLMs tailored for Romanian. We evaluate our methods on four different categories, including academic benchmarks, MT-Bench (manually translated), and a professionally built historical, cultural, and social benchmark adapted to Romanian. We argue for the usefulness and high performance of RoLLMs by obtaining state-of-the-art results across the board. We publicly release all resources (i.e., data, training and evaluation code, models) to support and encourage research on Romanian LLMs while concurrently creating a generalizable recipe, adequate for other low or less-resourced languages.

In recent years, Large Language Models (LLMs) have achieved almost human-like performance on various tasks. While some LLMs have been trained on multilingual data, most of the training data is in English. Hence, their performance in English greatly exceeds their performance in other languages. This document presents our approach to training and evaluating the first foundational and chat LLM specialized for Romanian.

Conspiracy theories have become a prominent and concerning aspect of online discourse, posing challenges to information integrity and societal trust. As such, we address conspiracy theory detection as proposed by the ACTI @ EVALITA 2023 shared task. The combination of pre-trained sentence Transformer models and data augmentation techniques enabled us to secure first place in the final leaderboard of both sub-tasks. Our methodology attained F1 scores of 85.71% in the binary classification and 91.23% for the fine-grained conspiracy topic classification, surpassing other competing systems.

Keyphrase identification and classification is a Natural Language Processing and Information Retrieval task that involves extracting relevant groups of words from a given text related to the main topic. In this work, we focus on extracting keyphrases from scientific documents. We introduce TA-DA, a Topic-Aware Domain Adaptation framework for keyphrase extraction that integrates Multi-Task Learning with Adversarial Training and Domain Adaptation. Our approach improves performance over baseline models by up to 5% in the exact match of the F1-score.

Detecting humor is a challenging task since words might share multiple valences and, depending on the context, the same words can be even used in offensive expressions. Neural network architectures based on Transformer obtain state-of-the-art results on several Natural Language Processing tasks, especially text classification. Adversarial learning, combined with other techniques such as multi-task learning, aids neural models learn the intrinsic properties of data. In this work, we describe our adversarial multi-task network, AMTL-Humor, used to detect and rate humor and offensive texts from Task 7 at SemEval-2021. Each branch from the model is focused on solving a related task, and consists of a BiLSTM layer followed by Capsule layers, on top of BERTweet used for generating contextualized embeddings. Our best model consists of an ensemble of all tested configurations, and achieves a 95.66% F1-score and 94.70% accuracy for Task 1a, while obtaining RMSE scores of 0.6200 and 0.5318 for Tasks 1b and 2, respectively.

Textual complexity is widely used to assess the difficulty of reading materials and writing quality in student essays. At a lexical level, word complexity can represent a building block for creating a comprehensive model of lexical networks that adequately estimates learners’ understanding. In order to best capture how lexical associations are created between related concepts, we propose automated indices of word complexity based on Age of Exposure (AoE). AOE indices computationally model the lexical learning process as a function of a learner's experience with language. This study describes a proof of concept based on the on a large-scale learning corpus (i.e., TASA). The results indicate that AoE indices yield strong associations with human ratings of age of acquisition, word frequency, entropy, and human lexical response latencies providing evidence of convergent validity.