Roberto Navigli

Sedrick Scott Keh, Rohit K. Bharadwaj, Emmy Liu et al. · cmu

We introduce EUREKA, an ensemble-based approach for performing automatic euphemism detection. We (1) identify and correct potentially mislabelled rows in the dataset, (2) curate an expanded corpus called EuphAug, (3) leverage model representations of Potentially Euphemistic Terms (PETs), and (4) explore using representations of semantically close sentences to aid in classification. Using our augmented dataset and kNN-based methods, EUREKA was able to achieve state-of-the-art results on the public leaderboard of the Euphemism Detection Shared Task, ranking first with a macro F1 score of 0.881. Our code is available at https://github.com/sedrickkeh/EUREKA.

Pere-Lluís Huguet Cabot, Simone Tedeschi, Axel-Cyrille Ngonga Ngomo et al.

Relation Extraction (RE) is a task that identifies relationships between entities in a text, enabling the acquisition of relational facts and bridging the gap between natural language and structured knowledge. However, current RE models often rely on small datasets with low coverage of relation types, particularly when working with languages other than English. In this paper, we address the above issue and provide two new resources that enable the training and evaluation of multilingual RE systems. First, we present SRED$^{\rm FM}$, an automatically annotated dataset covering 18 languages, 400 relation types, 13 entity types, totaling more than 40 million triplet instances. Second, we propose RED$^{\rm FM}$, a smaller, human-revised dataset for seven languages that allows for the evaluation of multilingual RE systems. To demonstrate the utility of these novel datasets, we experiment with the first end-to-end multilingual RE model, mREBEL, that extracts triplets, including entity types, in multiple languages. We release our resources and model checkpoints at https://www.github.com/babelscape/rebel

Simone Conia, Edoardo Barba, Alessandro Scirè et al.

One of the common traits of past and present approaches for Semantic Role Labeling (SRL) is that they rely upon discrete labels drawn from a predefined linguistic inventory to classify predicate senses and their arguments. However, we argue this need not be the case. In this paper, we present an approach that leverages Definition Modeling to introduce a generalized formulation of SRL as the task of describing predicate-argument structures using natural language definitions instead of discrete labels. Our novel formulation takes a first step towards placing interpretability and flexibility foremost, and yet our experiments and analyses on PropBank-style and FrameNet-style, dependency-based and span-based SRL also demonstrate that a flexible model with an interpretable output does not necessarily come at the expense of performance. We release our software for research purposes at https://github.com/SapienzaNLP/dsrl.

Luigi Procopio, Simone Conia, Edoardo Barba et al.

Local models have recently attained astounding performances in Entity Disambiguation (ED), with generative and extractive formulations being the most promising research directions. However, previous works limited their studies to using, as the textual representation of each candidate, only its Wikipedia title. Although certainly effective, this strategy presents a few critical issues, especially when titles are not sufficiently informative or distinguishable from one another. In this paper, we address this limitation and investigate to what extent more expressive textual representations can mitigate it. We thoroughly evaluate our approach against standard benchmarks in ED and find extractive formulations to be particularly well-suited to these representations: we report a new state of the art on 2 out of 6 benchmarks we consider and strongly improve the generalization capability over unseen patterns. We release our code, data and model checkpoints at https://github.com/SapienzaNLP/extend.

Giuliano Martinelli, Edoardo Barba, Roberto Navigli

Large autoregressive generative models have emerged as the cornerstone for achieving the highest performance across several Natural Language Processing tasks. However, the urge to attain superior results has, at times, led to the premature replacement of carefully designed task-specific approaches without exhaustive experimentation. The Coreference Resolution task is no exception; all recent state-of-the-art solutions adopt large generative autoregressive models that outperform encoder-based discriminative systems. In this work,we challenge this recent trend by introducing Maverick, a carefully designed - yet simple - pipeline, which enables running a state-of-the-art Coreference Resolution system within the constraints of an academic budget, outperforming models with up to 13 billion parameters with as few as 500 million parameters. Maverick achieves state-of-the-art performance on the CoNLL-2012 benchmark, training with up to 0.006x the memory resources and obtaining a 170x faster inference compared to previous state-of-the-art systems. We extensively validate the robustness of the Maverick framework with an array of diverse experiments, reporting improvements over prior systems in data-scarce, long-document, and out-of-domain settings. We release our code and models for research purposes at https://github.com/SapienzaNLP/maverick-coref.

Pavlo Vasylenko, Pere-Lluís Huguet Cabot, Abelardo Carlos Martínez Lorenzo et al.

Abstract Meaning Representation (AMR) is a Semantic Parsing formalism that aims at providing a semantic graph abstraction representing a given text. Current approaches are based on autoregressive language models such as BART or T5, fine-tuned through Teacher Forcing to obtain a linearized version of the AMR graph from a sentence. In this paper, we present LeakDistill, a model and method that explores a modification to the Transformer architecture, using structural adapters to explicitly incorporate graph information into the learned representations and improve AMR parsing performance. Our experiments show how, by employing word-to-node alignment to embed graph structural information into the encoder at training time, we can obtain state-of-the-art AMR parsing through self-knowledge distillation, even without the use of additional data. We release the code at \url{http://www.github.com/sapienzanlp/LeakDistill}.

Alessandro Scirè, Simone Conia, Simone Ciciliano et al.

In recent years, research in text summarization has mainly focused on the news domain, where texts are typically short and have strong layout features. The task of full-book summarization presents additional challenges which are hard to tackle with current resources, due to their limited size and availability in English only. To overcome these limitations, we present "Echoes from Alexandria", or in shortened form, "Echoes", a large resource for multilingual book summarization. Echoes features three novel datasets: i) Echo-Wiki, for multilingual book summarization, ii) Echo-XSum, for extremely-compressive multilingual book summarization, and iii) Echo-FairySum, for extractive book summarization. To the best of our knowledge, Echoes, with its thousands of books and summaries, is the largest resource, and the first to be multilingual, featuring 5 languages and 25 language pairs. In addition to Echoes, we also introduce a new extractive-then-abstractive baseline, and, supported by our experimental results and manual analysis of the summaries generated, we argue that this baseline is more suitable for book summarization than purely-abstractive approaches. We release our resource and software at https://github.com/Babelscape/echoes-from-alexandria in the hope of fostering innovative research in multilingual book summarization.

Abelardo Carlos Martínez Lorenzo, Pere-Lluís Huguet Cabot, Roberto Navigli

In this paper, we examine the current state-of-the-art in AMR parsing, which relies on ensemble strategies by merging multiple graph predictions. Our analysis reveals that the present models often violate AMR structural constraints. To address this issue, we develop a validation method, and show how ensemble models can exploit SMATCH metric weaknesses to obtain higher scores, but sometimes result in corrupted graphs. Additionally, we highlight the demanding need to compute the SMATCH score among all possible predictions. To overcome these challenges, we propose two novel ensemble strategies based on Transformer models, improving robustness to structural constraints, while also reducing the computational time. Our methods provide new insights for enhancing AMR parsers and metrics. Our code is available at \href{https://www.github.com/babelscape/AMRs-Assemble}{github.com/babelscape/AMRs-Assemble}.

Boshko Koloski, Senja Pollak, Roberto Navigli et al.

Large semantic knowledge bases are grounded in factual knowledge. However, recent approaches to dense text representations (i.e. embeddings) do not efficiently exploit these resources. Dense and robust representations of documents are essential for effectively solving downstream classification and retrieval tasks. This work demonstrates that injecting embedded information from knowledge bases can augment the performance of contemporary Large Language Model (LLM)-based representations for the task of text classification. Further, by considering automated machine learning (AutoML) with the fused representation space, we demonstrate it is possible to improve classification accuracy even if we use low-dimensional projections of the original representation space obtained via efficient matrix factorization. This result shows that significantly faster classifiers can be achieved with minimal or no loss in predictive performance, as demonstrated using five strong LLM baselines on six diverse real-life datasets. The code is freely available at \url{https://github.com/bkolosk1/bablfusion.git}.

Abelardo Carlos Martínez Lorenzo, Pere-Lluís Huguet Cabot, Roberto Navigli

This paper introduces a novel aligner for Abstract Meaning Representation (AMR) graphs that can scale cross-lingually, and is thus capable of aligning units and spans in sentences of different languages. Our approach leverages modern Transformer-based parsers, which inherently encode alignment information in their cross-attention weights, allowing us to extract this information during parsing. This eliminates the need for English-specific rules or the Expectation Maximization (EM) algorithm that have been used in previous approaches. In addition, we propose a guided supervised method using alignment to further enhance the performance of our aligner. We achieve state-of-the-art results in the benchmarks for AMR alignment and demonstrate our aligner's ability to obtain them across multiple languages. Our code will be available at \href{https://www.github.com/Babelscape/AMR-alignment}{github.com/Babelscape/AMR-alignment}.

Vera Provatorova, Simone Tedeschi, Svitlana Vakulenko et al.

Entity disambiguation (ED) is the task of mapping an ambiguous entity mention to the corresponding entry in a structured knowledge base. Previous research showed that entity overshadowing is a significant challenge for existing ED models: when presented with an ambiguous entity mention, the models are much more likely to rank a more frequent yet less contextually relevant entity at the top. Here, we present NICE, an iterative approach that uses entity type information to leverage context and avoid over-relying on the frequency-based prior. Our experiments show that NICE achieves the best performance results on the overshadowed entities while still performing competitively on the frequent entities.

Riccardo Orlando, Pere-Lluis Huguet Cabot, Edoardo Barba et al.

Entity Linking (EL) and Relation Extraction (RE) are fundamental tasks in Natural Language Processing, serving as critical components in a wide range of applications. In this paper, we propose ReLiK, a Retriever-Reader architecture for both EL and RE, where, given an input text, the Retriever module undertakes the identification of candidate entities or relations that could potentially appear within the text. Subsequently, the Reader module is tasked to discern the pertinent retrieved entities or relations and establish their alignment with the corresponding textual spans. Notably, we put forward an innovative input representation that incorporates the candidate entities or relations alongside the text, making it possible to link entities or extract relations in a single forward pass and to fully leverage pre-trained language models contextualization capabilities, in contrast with previous Retriever-Reader-based methods, which require a forward pass for each candidate. Our formulation of EL and RE achieves state-of-the-art performance in both in-domain and out-of-domain benchmarks while using academic budget training and with up to 40x inference speed compared to competitors. Finally, we show how our architecture can be used seamlessly for Information Extraction (cIE), i.e. EL + RE, and setting a new state of the art by employing a shared Reader that simultaneously extracts entities and relations.

Stefano Perrella, Lorenzo Proietti, Alessandro Scirè et al.

Annually, at the Conference of Machine Translation (WMT), the Metrics Shared Task organizers conduct the meta-evaluation of Machine Translation (MT) metrics, ranking them according to their correlation with human judgments. Their results guide researchers toward enhancing the next generation of metrics and MT systems. With the recent introduction of neural metrics, the field has witnessed notable advancements. Nevertheless, the inherent opacity of these metrics has posed substantial challenges to the meta-evaluation process. This work highlights two issues with the meta-evaluation framework currently employed in WMT, and assesses their impact on the metrics rankings. To do this, we introduce the concept of sentinel metrics, which are designed explicitly to scrutinize the meta-evaluation process's accuracy, robustness, and fairness. By employing sentinel metrics, we aim to validate our findings, and shed light on and monitor the potential biases or inconsistencies in the rankings. We discover that the present meta-evaluation framework favors two categories of metrics: i) those explicitly trained to mimic human quality assessments, and ii) continuous metrics. Finally, we raise concerns regarding the evaluation capabilities of state-of-the-art metrics, emphasizing that they might be basing their assessments on spurious correlations found in their training data.

Vivek Iyer, Edoardo Barba, Alexandra Birch et al.

Lexical ambiguity is a significant and pervasive challenge in Neural Machine Translation (NMT), with many state-of-the-art (SOTA) NMT systems struggling to handle polysemous words (Campolungo et al., 2022). The same holds for the NMT pretraining paradigm of denoising synthetic "code-switched" text (Pan et al., 2021; Iyer et al., 2023), where word senses are ignored in the noising stage -- leading to harmful sense biases in the pretraining data that are subsequently inherited by the resulting models. In this work, we introduce Word Sense Pretraining for Neural Machine Translation (WSP-NMT) - an end-to-end approach for pretraining multilingual NMT models leveraging word sense-specific information from Knowledge Bases. Our experiments show significant improvements in overall translation quality. Then, we show the robustness of our approach to scale to various challenging data and resource-scarce scenarios and, finally, report fine-grained accuracy improvements on the DiBiMT disambiguation benchmark. Our studies yield interesting and novel insights into the merits and challenges of integrating word sense information and structured knowledge in multilingual pretraining for NMT.

Riccardo Orlando, Simone Conia, Roberto Navigli

Although we have witnessed impressive progress in Semantic Role Labeling (SRL), most of the research in the area is carried out assuming that the majority of predicates are verbs. Conversely, predicates can also be expressed using other parts of speech, e.g., nouns and adjectives. However, non-verbal predicates appear in the benchmarks we commonly use to measure progress in SRL less frequently than in some real-world settings -- newspaper headlines, dialogues, and tweets, among others. In this paper, we put forward a new PropBank dataset which boasts wide coverage of multiple predicate types. Thanks to it, we demonstrate empirically that standard benchmarks do not provide an accurate picture of the current situation in SRL and that state-of-the-art systems are still incapable of transferring knowledge across different predicate types. Having observed these issues, we also present a novel, manually-annotated challenge set designed to give equal importance to verbal, nominal, and adjectival predicate-argument structures. We use such dataset to investigate whether we can leverage different linguistic resources to promote knowledge transfer. In conclusion, we claim that SRL is far from "solved", and its integration with other semantic tasks might enable significant improvements in the future, especially for the long tail of non-verbal predicates, thereby facilitating further research on SRL for non-verbal predicates.

Alessandro Scirè, Karim Ghonim, Roberto Navigli

Recent advancements in text summarization, particularly with the advent of Large Language Models (LLMs), have shown remarkable performance. However, a notable challenge persists as a substantial number of automatically-generated summaries exhibit factual inconsistencies, such as hallucinations. In response to this issue, various approaches for the evaluation of consistency for summarization have emerged. Yet, these newly-introduced metrics face several limitations, including lack of interpretability, focus on short document summaries (e.g., news articles), and computational impracticality, especially for LLM-based metrics. To address these shortcomings, we propose Factuality Evaluation of summarization based on Natural language Inference and Claim Extraction (FENICE), a more interpretable and efficient factuality-oriented metric. FENICE leverages an NLI-based alignment between information in the source document and a set of atomic facts, referred to as claims, extracted from the summary. Our metric sets a new state of the art on AGGREFACT, the de-facto benchmark for factuality evaluation. Moreover, we extend our evaluation to a more challenging setting by conducting a human annotation process of long-form summarization. In the hope of fostering research in summarization factuality evaluation, we release the code of our metric and our factuality annotations of long-form summarization at https://github.com/Babelscape/FENICE.

Taishi Nakamura, Mayank Mishra, Simone Tedeschi et al. · ibm-research, stanford

Pretrained language models are an integral part of AI applications, but their high computational cost for training limits accessibility. Initiatives such as Bloom and StarCoder aim to democratize access to pretrained models for collaborative community development. Despite these efforts, such models encounter challenges such as limited multilingual capabilities, risks of catastrophic forgetting during continual pretraining, and the high costs of training models from scratch, alongside the need to align with AI safety standards and regulatory frameworks. This paper presents Aurora-M, a 15B parameter multilingual open-source model trained on English, Finnish, Hindi, Japanese, Vietnamese, and code. Continually pretrained from StarCoderPlus on 435B additional tokens, Aurora-M surpasses 2T tokens in total training token count. It is the first open-source multilingual model fine-tuned on human-reviewed safety instructions, thus aligning its development not only with conventional red-teaming considerations, but also with the specific concerns articulated in the Biden-Harris Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence. We evaluate Aurora-M across a wide range of tasks and languages, showcasing its robustness against catastrophic forgetting and its superior performance in multilingual settings, particularly in safety evaluations. We open-source Aurora-M and its variants to encourage responsible open-source development of large language models at https://huggingface.co/aurora-m.

Giuliano Martinelli, Tommaso Bonomo, Pere-Lluís Huguet Cabot et al.

Coreference Resolution systems are typically evaluated on benchmarks containing small- to medium-scale documents. When it comes to evaluating long texts, however, existing benchmarks, such as LitBank, remain limited in length and do not adequately assess system capabilities at the book scale, i.e., when co-referring mentions span hundreds of thousands of tokens. To fill this gap, we first put forward a novel automatic pipeline that produces high-quality Coreference Resolution annotations on full narrative texts. Then, we adopt this pipeline to create the first book-scale coreference benchmark, BOOKCOREF, with an average document length of more than 200,000 tokens. We carry out a series of experiments showing the robustness of our automatic procedure and demonstrating the value of our resource, which enables current long-document coreference systems to gain up to +20 CoNLL-F1 points when evaluated on full books. Moreover, we report on the new challenges introduced by this unprecedented book-scale setting, highlighting that current models fail to deliver the same performance they achieve on smaller documents. We release our data and code to encourage research and development of new book-scale Coreference Resolution systems at https://github.com/sapienzanlp/bookcoref.

Raffaele Pisano, Roberto Navigli

Process Reward Models (PRMs) have emerged as a powerful tool for providing step-level feedback when evaluating the reasoning of Large Language Models (LLMs), which frequently produce chains of thought (CoTs) containing errors even when the final answer is correct. However, existing PRM datasets remain expensive to construct, prone to annotation errors, and predominantly limited to the mathematical domain. This work introduces a novel and scalable approach to PRM dataset generation based on planning logical problems expressed in the Planning Domain Definition Language (PDDL). Using this method, we generate a corpus of approximately one million reasoning steps across various PDDL domains and use it to train PRMs. Experimental results show that augmenting widely-used PRM training datasets with PDDL-derived data yields substantial improvements in both mathematical and non-mathematical reasoning, as demonstrated across multiple benchmarks. These findings indicate that planning problems constitute a scalable and effective resource for generating robust, precise, and fine-grained training data for PRMs, going beyond the classical mathematical sources that dominate this field.

Bruno Gatti, Giuliano Martinelli, Roberto Navigli

Coreference resolution is typically evaluated using aggregate statistical metrics such as CoNLL-F1, which measure structural overlap between predicted and gold clusters. While widely used, these metrics offer limited diagnostic insights, penalizing errors without revealing whether a system struggles with specific semantic categories, such as people, locations, or events, and making it difficult to interpret model capabilities or derive actionable improvements. We address this gap by introducing a semantically-enhanced evaluation framework for coreference resolution. Our approach overlays Concept and Named Entity Recognition (CNER) onto coreference outputs, assigning semantic labels to nominal mentions and propagating them to entire coreference clusters. This enables the computation of typed scores aimed at evaluating mention extraction and linking capabilities stratified by semantic class. Across our experiments on OntoNotes, LitBank, and PreCo, we show that our framework uncovers systematic weaknesses that remain obscured by aggregate metrics. Furthermore, we demonstrate that these diagnostics can be used to design targeted, low-cost data augmentation strategies, achieving measurable out-of-domain improvements.

Tommaso Bonomo, Luca Gioffré, Roberto Navigli

Question Answering (QA) on narrative text poses a unique challenge to current systems, requiring a deep understanding of long, complex documents. However, the reliability of NarrativeQA, the most widely used benchmark in this domain, is hindered by noisy documents and flawed QA pairs. In this work, we introduce LiteraryQA, a high-quality subset of NarrativeQA focused on literary works. Using a human- and LLM-validated pipeline, we identify and correct low-quality QA samples while removing extraneous text from source documents. We then carry out a meta-evaluation of automatic metrics to clarify how systems should be evaluated on LiteraryQA. This analysis reveals that all n-gram-based metrics have a low system-level correlation to human judgment, while LLM-as-a-Judge evaluations, even with small open-weight models, can strongly agree with the ranking identified by humans. Finally, we benchmark a set of long-context LLMs on LiteraryQA. We release our code and data at https://github.com/SapienzaNLP/LiteraryQA.

Simone Tedeschi, Felix Friedrich, Patrick Schramowski et al.

When building Large Language Models (LLMs), it is paramount to bear safety in mind and protect them with guardrails. Indeed, LLMs should never generate content promoting or normalizing harmful, illegal, or unethical behavior that may contribute to harm to individuals or society. This principle applies to both normal and adversarial use. In response, we introduce ALERT, a large-scale benchmark to assess safety based on a novel fine-grained risk taxonomy. It is designed to evaluate the safety of LLMs through red teaming methodologies and consists of more than 45k instructions categorized using our novel taxonomy. By subjecting LLMs to adversarial testing scenarios, ALERT aims to identify vulnerabilities, inform improvements, and enhance the overall safety of the language models. Furthermore, the fine-grained taxonomy enables researchers to perform an in-depth evaluation that also helps one to assess the alignment with various policies. In our experiments, we extensively evaluate 10 popular open- and closed-source LLMs and demonstrate that many of them still struggle to attain reasonable levels of safety.

Francesco Maria Molfese, Luca Moroni, Luca Gioffré et al.

One of the most widely used tasks for evaluating Large Language Models (LLMs) is Multiple-Choice Question Answering (MCQA). While open-ended question answering tasks are more challenging to evaluate, MCQA tasks are, in principle, easier to assess, as the model's answer is thought to be simple to extract and is compared directly to a set of predefined choices. However, recent studies have started to question the reliability of MCQA evaluation, showing that multiple factors can significantly impact the reported performance of LLMs, especially when the model generates free-form text before selecting one of the answer choices. In this work, we shed light on the inconsistencies of MCQA evaluation strategies, which can lead to inaccurate and misleading model comparisons. We systematically analyze whether existing answer extraction methods are aligned with human judgment, and how they are influenced by answer constraints in the prompt across different domains. Our experiments demonstrate that traditional evaluation strategies often underestimate LLM capabilities, while LLM-based answer extractors are prone to systematic errors. Moreover, we reveal a fundamental trade-off between including format constraints in the prompt to simplify answer extraction and allowing models to generate free-form text to improve reasoning. Our findings call for standardized evaluation methodologies and highlight the need for more reliable and consistent MCQA evaluation practices.

Alessandro Scirè, Andrei Stefan Bejgu, Simone Tedeschi et al.

After the introduction of Large Language Models (LLMs), there have been substantial improvements in the performance of Natural Language Generation (NLG) tasks, including Text Summarization and Machine Translation. However, LLMs still produce outputs containing hallucinations, that is, content not grounded in factual information. Therefore, developing methods to assess the factuality of LLMs has become urgent. Indeed, resources for factuality evaluation have recently emerged. Although challenging, these resources face one or more of the following limitations: (i) they are tailored to a specific task or domain; (ii) they are limited in size, thereby preventing the training of new factuality evaluators; (iii) they are designed for simpler verification tasks, such as claim verification. To address these issues, we introduce LLM-Oasis, to the best of our knowledge the largest resource for training end-to-end factuality evaluators. LLM-Oasis is constructed by extracting claims from Wikipedia, falsifying a subset of these claims, and generating pairs of factual and unfactual texts. We then rely on human annotators to both validate the quality of our dataset and to create a gold standard test set for benchmarking factuality evaluation systems. Our experiments demonstrate that LLM-Oasis presents a significant challenge for state-of-the-art LLMs, with GPT-4o achieving up to 60% accuracy in our proposed end-to-end factuality evaluation task, highlighting its potential to drive future research in the field.

Felix Friedrich, Simone Tedeschi, Patrick Schramowski et al.

Building safe Large Language Models (LLMs) across multiple languages is essential in ensuring both safe access and linguistic diversity. To this end, we conduct a large-scale, comprehensive safety evaluation of the current LLM landscape. For this purpose, we introduce M-ALERT, a multilingual benchmark that evaluates the safety of LLMs in five languages: English, French, German, Italian, and Spanish. M-ALERT includes 15k high-quality prompts per language, totaling 75k, with category-wise annotations. Our extensive experiments on 39 state-of-the-art LLMs highlight the importance of language-specific safety analysis, revealing that models often exhibit significant inconsistencies in safety across languages and categories. For instance, Llama3.2 shows high unsafety in category crime_tax for Italian but remains safe in other languages. Similar inconsistencies can be observed across all models. In contrast, certain categories, such as substance_cannabis and crime_propaganda, consistently trigger unsafe responses across models and languages. These findings underscore the need for robust multilingual safety practices in LLMs to ensure responsible usage across diverse communities.

Lorenzo Proietti, Stefano Perrella, Vilém Zouhar et al.

Machine translation quality has steadily improved over the years, achieving near-perfect translations in recent benchmarks. These high-quality outputs make it difficult to distinguish between state-of-the-art models and to identify areas for future improvement. In this context, automatically identifying texts where machine translation systems struggle holds promise for developing more discriminative evaluations and guiding future research. In this work, we address this gap by formalizing the task of translation difficulty estimation, defining a text's difficulty based on the expected quality of its translations. We introduce a new metric to evaluate difficulty estimators and use it to assess both baselines and novel approaches. Finally, we demonstrate the practical utility of difficulty estimators by using them to construct more challenging benchmarks for machine translation. Our results show that dedicated models outperform both heuristic-based methods and LLM-as-a-judge approaches, with Sentinel-src achieving the best performance. Thus, we release two improved models for difficulty estimation, Sentinel-src-24 and Sentinel-src-25, which can be used to scan large collections of texts and select those most likely to challenge contemporary machine translation systems.

Lorenzo Proietti, Stefano Perrella, Roberto Navigli

In Machine Translation (MT) evaluation, metric performance is assessed based on agreement with human judgments. In recent years, automatic metrics have demonstrated increasingly high levels of agreement with humans. To gain a clearer understanding of metric performance and establish an upper bound, we incorporate human baselines in the MT meta-evaluation, that is, the assessment of MT metrics' capabilities. Our results show that human annotators are not consistently superior to automatic metrics, with state-of-the-art metrics often ranking on par with or higher than human baselines. Despite these findings suggesting human parity, we discuss several reasons for caution. Finally, we explore the broader implications of our results for the research field, asking: Can we still reliably measure improvements in MT evaluation? With this work, we aim to shed light on the limits of our ability to measure progress in the field, fostering discussion on an issue that we believe is crucial to the entire MT evaluation community.

Luca Moroni, Giovanni Puccetti, Pere-Lluis Huguet Cabot et al.

The number of pretrained Large Language Models (LLMs) is increasing steadily, though the majority are designed predominantly for the English language. While state-of-the-art LLMs can handle other languages, due to language contamination or some degree of multilingual pretraining data, they are not optimized for non-English languages, leading to inefficient encoding (high token "fertility") and slower inference speed. In this work, we thoroughly compare a variety of vocabulary adaptation techniques for optimizing English LLMs for the Italian language, and put forward Semantic Alignment Vocabulary Adaptation (SAVA), a novel method that leverages neural mapping for vocabulary substitution. SAVA achieves competitive performance across multiple downstream tasks, enhancing grounded alignment strategies. We adapt two LLMs: Mistral-7b-v0.1, reducing token fertility by 25\%, and Llama-3.1-8B, optimizing the vocabulary and reducing the number of parameters by 1 billion. We show that, following the adaptation of the vocabulary, these models can recover their performance with a relatively limited stage of continual training on the target language. Finally, we test the capabilities of the adapted models on various multi-choice and generative tasks.

Andrei Stefan Bejgu, Edoardo Barba, Luigi Procopio et al.

Word Sense Disambiguation (WSD) is the task of associating a word in a given context with its most suitable meaning among a set of possible candidates. While the task has recently witnessed renewed interest, with systems achieving performances above the estimated inter-annotator agreement, at the time of writing it still struggles to find downstream applications. We argue that one of the reasons behind this is the difficulty of applying WSD to plain text. Indeed, in the standard formulation, models work under the assumptions that a) all the spans to disambiguate have already been identified, and b) all the possible candidate senses of each span are provided, both of which are requirements that are far from trivial. In this work, we present a new task called Word Sense Linking (WSL) where, given an input text and a reference sense inventory, systems have to both identify which spans to disambiguate and then link them to their most suitable meaning.We put forward a transformer-based architecture for the task and thoroughly evaluate both its performance and those of state-of-the-art WSD systems scaled to WSL, iteratively relaxing the assumptions of WSD. We hope that our work will foster easier integration of lexical semantics into downstream applications.

Francesco Maria Molfese, Luca Moroni, Ciro Porcaro et al.

While Small Language Models (SLMs) have demonstrated promising performance on an increasingly wide array of commonsense reasoning benchmarks, current evaluation practices rely almost exclusively on the accuracy of their final answers, neglecting the validity of the reasoning processes that lead to those answers. To address this issue, we introduce ReTraceQA, a novel benchmark that introduces process-level evaluation for commonsense reasoning tasks. Our expert-annotated dataset reveals that in a substantial portion of instances (14-24%), SLMs provide correct final answers despite flawed reasoning processes, suggesting that the capabilities of SLMs are often overestimated by evaluation metrics that focus only on comparing the final answer with the ground truth. Indeed, we show that when employing strong Large Language Models (LLMs) as automated judges for reasoning-aware evaluation rather than answer-only metrics, SLM performance drops significantly across all models and datasets, with scores decreasing by up to 25%.

Domenico Meconi, Simone Stirpe, Federico Martelli et al.

Understanding the meaning of words in context is a fundamental capability for Large Language Models (LLMs). Despite extensive evaluation efforts, the extent to which LLMs show evidence that they truly grasp word senses remains underexplored. In this paper, we address this gap by evaluating both i) the Word Sense Disambiguation (WSD) capabilities of instruction-tuned LLMs, comparing their performance to state-of-the-art systems specifically designed for the task, and ii) the ability of two top-performing open- and closed-source LLMs to understand word senses in three generative settings: definition generation, free-form explanation, and example generation. Notably, we find that, in the WSD task, leading models such as GPT-4o and DeepSeek-V3 achieve performance on par with specialized WSD systems, while also demonstrating greater robustness across domains and levels of difficulty. In the generation tasks, results reveal that LLMs can explain the meaning of words in context up to 98\% accuracy, with the highest performance observed in the free-form explanation task, which best aligns with their generative capabilities.

Boshko Koloski, Senja Pollak, Roberto Navigli et al.

Building on the success of Large Language Models (LLMs), LLM-based representations have dominated the document representation landscape, achieving great performance on the document embedding benchmarks. However, the high-dimensional, computationally expensive embeddings from LLMs tend to be either too generic or inefficient for domain-specific applications. To address these limitations, we introduce FuDoBa a Bayesian optimisation-based method that integrates LLM-based embeddings with domain-specific structured knowledge, sourced both locally and from external repositories like WikiData. This fusion produces low-dimensional, task-relevant representations while reducing training complexity and yielding interpretable early-fusion weights for enhanced classification performance. We demonstrate the effectiveness of our approach on six datasets in two domains, showing that when paired with robust AutoML-based classifiers, our proposed representation learning approach performs on par with, or surpasses, those produced solely by the proprietary LLM-based embedding baselines.

Klim Zaporojets, Daniel Daza, Edoardo Barba et al.

Knowledge Graphs (KGs) are structured knowledge repositories containing entities and relations between them. In this paper, we investigate the problem of automatically updating KGs over time with respect to the evolution of knowledge in unstructured textual sources. This problem requires identifying a wide range of update operations based on the state of an existing KG at a specific point in time. This contrasts with traditional information extraction pipelines, which extract knowledge from text independently of the current state of a KG. To address this challenge, we propose a method for lifelong construction of a dataset consisting of Wikidata KG snapshots over time and Wikipedia passages paired with the corresponding edit operations that they induce in a particular KG snapshot. The resulting dataset comprises 376K Wikipedia passages aligned with a total of 1.25M KG edits over 10 different snapshots of Wikidata from 2019 to 2025. Our experimental results highlight challenges in updating KG snapshots based on emerging textual knowledge, positioning the dataset as a valuable benchmark for future research. We will publicly release our dataset and model implementations.

Simone Tedeschi, Johan Bos, Thierry Declerck et al.

In the last five years, there has been a significant focus in Natural Language Processing (NLP) on developing larger Pretrained Language Models (PLMs) and introducing benchmarks such as SuperGLUE and SQuAD to measure their abilities in language understanding, reasoning, and reading comprehension. These PLMs have achieved impressive results on these benchmarks, even surpassing human performance in some cases. This has led to claims of superhuman capabilities and the provocative idea that certain tasks have been solved. In this position paper, we take a critical look at these claims and ask whether PLMs truly have superhuman abilities and what the current benchmarks are really evaluating. We show that these benchmarks have serious limitations affecting the comparison between humans and PLMs and provide recommendations for fairer and more transparent benchmarks.

Aidan Hogan, Eva Blomqvist, Michael Cochez et al.

In this paper we provide a comprehensive introduction to knowledge graphs, which have recently garnered significant attention from both industry and academia in scenarios that require exploiting diverse, dynamic, large-scale collections of data. After some opening remarks, we motivate and contrast various graph-based data models and query languages that are used for knowledge graphs. We discuss the roles of schema, identity, and context in knowledge graphs. We explain how knowledge can be represented and extracted using a combination of deductive and inductive techniques. We summarise methods for the creation, enrichment, quality assessment, refinement, and publication of knowledge graphs. We provide an overview of prominent open knowledge graphs and enterprise knowledge graphs, their applications, and how they use the aforementioned techniques. We conclude with high-level future research directions for knowledge graphs.

Tommaso Pasini, Francesco Maria Elia, Roberto Navigli

We release to the community six large-scale sense-annotated datasets in multiple language to pave the way for supervised multilingual Word Sense Disambiguation. Our datasets cover all the nouns in the English WordNet and their translations in other languages for a total of millions of sense-tagged sentences. Experiments prove that these corpora can be effectively used as training sets for supervised WSD systems, surpassing the state of the art for low-resourced languages and providing competitive results for English, where manually annotated training sets are accessible. The data is available at trainomatic.org.

Mohammad Taher Pilehvar, Jose Camacho-Collados, Roberto Navigli et al.

Lexical ambiguity can impede NLP systems from accurate understanding of semantics. Despite its potential benefits, the integration of sense-level information into NLP systems has remained understudied. By incorporating a novel disambiguation algorithm into a state-of-the-art classification model, we create a pipeline to integrate sense-level information into downstream NLP applications. We show that a simple disambiguation of the input text can lead to consistent performance improvement on multiple topic categorization and polarity detection datasets, particularly when the fine granularity of the underlying sense inventory is reduced and the document is sufficiently large. Our results also point to the need for sense representation research to focus more on in vivo evaluations which target the performance in downstream NLP applications rather than artificial benchmarks.

Massimiliano Mancini, Jose Camacho-Collados, Ignacio Iacobacci et al.

Word embeddings are widely used in Natural Language Processing, mainly due to their success in capturing semantic information from massive corpora. However, their creation process does not allow the different meanings of a word to be automatically separated, as it conflates them into a single vector. We address this issue by proposing a new model which learns word and sense embeddings jointly. Our model exploits large corpora and knowledge from semantic networks in order to produce a unified vector space of word and sense embeddings. We evaluate the main features of our approach both qualitatively and quantitatively in a variety of tasks, highlighting the advantages of the proposed method in comparison to state-of-the-art word- and sense-based models.

José Camacho Collados, Claudio Delli Bovi, Alessandro Raganato et al.

Linking concepts and named entities to knowledge bases has become a crucial Natural Language Understanding task. In this respect, recent works have shown the key advantage of exploiting textual definitions in various Natural Language Processing applications. However, to date there are no reliable large-scale corpora of sense-annotated textual definitions available to the research community. In this paper we present a large-scale high-quality corpus of disambiguated glosses in multiple languages, comprising sense annotations of both concepts and named entities from a unified sense inventory. Our approach for the construction and disambiguation of the corpus builds upon the structure of a large multilingual semantic network and a state-of-the-art disambiguation system; first, we gather complementary information of equivalent definitions across different languages to provide context for disambiguation, and then we combine it with a semantic similarity-based refinement. As a result we obtain a multilingual corpus of textual definitions featuring over 38 million definitions in 263 languages, and we make it freely available at http://lcl.uniroma1.it/disambiguated-glosses. Experiments on Open Information Extraction and Sense Clustering show how two state-of-the-art approaches improve their performance by integrating our disambiguated corpus into their pipeline.

José Camacho-Collados, Ignacio Iacobacci, Roberto Navigli et al.

Representing the semantics of linguistic items in a machine-interpretable form has been a major goal of Natural Language Processing since its earliest days. Among the range of different linguistic items, words have attracted the most research attention. However, word representations have an important limitation: they conflate different meanings of a word into a single vector. Representations of word senses have the potential to overcome this inherent limitation. Indeed, the representation of individual word senses and concepts has recently gained in popularity with several experimental results showing that a considerable performance improvement can be achieved across different NLP applications upon moving from word level to the deeper sense and concept levels. Another interesting point regarding the representation of concepts and word senses is that these models can be seamlessly applied to other linguistic items, such as words, phrases and sentences.

Tiziano Flati, Roberto Navigli

Bilingual machine-readable dictionaries are knowledge resources useful in many automatic tasks. However, compared to monolingual computational lexicons like WordNet, bilingual dictionaries typically provide a lower amount of structured information, such as lexical and semantic relations, and often do not cover the entire range of possible translations for a word of interest. In this paper we present Cycles and Quasi-Cycles (CQC), a novel algorithm for the automated disambiguation of ambiguous translations in the lexical entries of a bilingual machine-readable dictionary. The dictionary is represented as a graph, and cyclic patterns are sought in the graph to assign an appropriate sense tag to each translation in a lexical entry. Further, we use the algorithms output to improve the quality of the dictionary itself, by suggesting accurate solutions to structural problems such as misalignments, partial alignments and missing entries. Finally, we successfully apply CQC to the task of synonym extraction.