Research

Heuristic-informed mixture of experts for link prediction in multilayer networks

Lucio La Cava, Domenico Mandaglio, Lorenzo Zangari, Andrea Tagarelli

Information Sciences · 23 Jan 2026 · doi:10.1016/j.ins.2026.123106

Link prediction algorithms for multilayer networks are in principle required to effectively account for the entire layered structure while capturing the unique contexts offered by each layer. However, many existing approaches excel at predicting specific links in certain layers but struggle with others, as they fail to effectively leverage the diverse information encoded across different network layers. In this paper, we present MoE-ML-LP, the first Mixture-of-Experts (MoE) framework specifically designed for multilayer link prediction. Building on top of multilayer heuristics for link prediction, MoE-ML-LP synthesizes the decisions taken by diverse experts, resulting in significantly enhanced predictive capabilities. Our extensive experimental evaluation on real-world and synthetic networks demonstrates that MoE-ML-LP consistently outperforms several baselines and competing methods, achieving remarkable improvements of +60% in Mean Reciprocal Rank, +82% in Hits@1, +55% in Hits@5, and +41% in Hits@10. Furthermore, MoE-ML-LP features a modular architecture that enables the seamless integration of newly developed experts without necessitating the re-training of the entire framework, fostering efficiency and scalability to new experts, and paving the way for future advancements in link prediction.

Polarized Communities Meet Densest Subgraph: Efficient and Effective Polarization Detection in Signed Networks

Francesco Gullo, Domenico Mandaglio, Andrea Tagarelli

ACM Transactions on Knowledge Discovery from Data · 13 Jan 2026 · doi:10.1145/3779064

Signed networks represent interactions among users (nodes), with edges labeled as positive for friendly relations and negative for antagonistic ones. The 2-Polarized-Communities (2pc) combinatorial optimization problem seeks two disjoint polarized communities in a signed network, so as to satisfy three conditions - most edges within each community are positive, most edges between communities are negative, and the number of edges satisfying these conditions is high compared to the number of nodes in the communities. The Densest Subgraph (ds) problem in unsigned networks consists in finding a subgraph that exhibits maximum ratio between number of edges and number of nodes. Although the 2pc problem intuitively suggests finding a dense subgraph, no prior work has explored the implicitly optimized density measure or algorithmic methods from the rich, yet distinct, literature on the ds problem (in unsigned networks) and applied them to 2pc. This work bridges this gap by formally establishing a link between the two problems and introducing a highly efficient and effective greedy algorithm inspired by ds methods to solve 2pc. Experimental results on synthetic and real datasets demonstrate the superior performance of our method compared to competing approaches in terms of both accuracy and efficiency.

OpenTuringBench: An Open-Model-based Benchmark and Framework for Machine-Generated Text Detection and Attribution

Lucio La Cava, Andrea Tagarelli

Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing · 23 Nov 2025 · doi:10.18653/v1/2025.emnlp-main.1354

Open Large Language Models (OLLMs) are increasingly leveraged in generative AI applications, posing new challenges for detecting their outputs. We propose OpenTuringBench, a new benchmark based on OLLMs, designed to train and evaluate machine-generated text detectors on the Turing Test and Authorship Attribution problems. OpenTuringBench focuses on a representative set of OLLMs, and features a number of challenging evaluation tasks, including human/machine-manipulated texts, out-of-domain texts, and texts from previously unseen models. We also provide OTBDetector, a contrastive learning framework to detect and attribute OLLM-based machine-generated texts. Results highlight the relevance and varying degrees of difficulty of the OpenTuringBench tasks, with our detector achieving remarkable capabilities across the various tasks and outperforming most existing detectors.

E2MoCase: A Dataset for Emotional, Event and Moral Observations in News Articles on High-impact Legal Cases

Candida M. Greco, Lorenzo Zangari, Davide Picca, Andrea Tagarelli

Proceedings of the 34th ACM International Conference on Information and Knowledge Management · 20 Nov 2025 · doi:10.1145/3746252.3761618

The way the media report on legal cases can significantly shape public opinion, often embedding subtle biases that influence societal views on justice, fairness, and morality. Analyzing these narratives requires a holistic approach that captures their emotional tone, moral framing, and the specific events they convey. In this work, we introduce E2MoCase, a novel dataset that enables integrated analysis of emotions, morality, and events within legal narratives and media coverage. We leverage NLP models to extract events and predict morality and emotions, providing a multidimensional perspective on how legal cases are portrayed in news articles. Our experimental evaluation showed that E2MoCase is beneficial for addressing emotion- and morality-based tasks, which is also confirmed by a human evaluation of the annotations.

PersonaGen: A Persona-Driven Open-Ended Machine-Generated Text Dataset

Carmelo Gugliotta, Lucio La Cava, Andrea Tagarelli

Proceedings of the 34th ACM International Conference on Information and Knowledge Management · 20 Nov 2025 · doi:10.1145/3746252.3761611

We present PersonaGen, a novel dataset for investigating persona-driven machine-generated text (MGT) produced by Open Large Language Models (OLLMS). PersonaGen is specifically designed to investigate how synthetic persona profiles affect, guide, or manifest in MGT. We built PersonaGen by pairing curated persona-profiles (i.e., description of characteristics, background, and goals) across eight thematic domains (e.g., Physics, Education, Medicine) with prompts covering various narrative or opinion-style content (e.g., stories, commonsense). Open-ended generations were produced by six representative OLLMs, yielding a total of 1.44 million persona-driven generations. PersonaGen supports multiple research tasks, such as machine-generated text attribution, persona category detection, and persona profile identification, thus providing a valuable resource for studying LLM controllability and role-playing behavior, as well as the impact of persona profile conditioning in downstream tasks. We have released PersonaGen on the Hugging Face platform at https://doi.org/10.57967/hf/5805.

Machines in the Crowd? Measuring the Footprint of Machine-Generated Text on Reddit

Lucio La Cava, Luca Maria Aiello, Andrea Tagarelli

arXiv · 08 Oct 2025 · doi:10.48550/arXiv.2510.07226

Generative Artificial Intelligence is reshaping online communication by enabling large-scale production of Machine-Generated Text (MGT) at low cost. While its presence is rapidly growing across the Web, little is known about how MGT integrates into social media environments. In this paper, we present the first large-scale characterization of MGT on Reddit. Using a state-of-the-art statistical method for detection of MGT, we analyze over two years of activity (2022-2024) across 51 subreddits representative of Reddit’s main community types such as information seeking, social support, and discussion. We study the concentration of MGT across communities and over time, and compared MGT to human-authored text in terms of social signals it expresses and engagement it receives. Our very conservative estimate of MGT prevalence indicates that synthetic text is marginally present on Reddit, but it can reach peaks of up to 9% in some communities in some months. MGT is unevenly distributed across communities, more prevalent in subreddits focused on technical knowledge and social support, and often concentrated in the activity of a small fraction of users. MGT also conveys distinct social signals of warmth and status giving typical of language of AI assistants. Despite these stylistic differences, MGT achieves engagement levels comparable than human-authored content and in a few cases even higher, suggesting that AI-generated text is becoming an organic component of online social discourse. This work offers the first perspective on the MGT footprint on Reddit, paving the way for new investigations involving platform governance, detection strategies, and community dynamics.

Toward Preference-aligned Large Language Models via Residual-based Model Steering

Lucio La Cava, Andrea Tagarelli

arXiv · 28 Sep 2025 · doi:10.48550/arXiv.2509.23982

Preference alignment is a critical step in making Large Language Models (LLMs) useful and aligned with (human) preferences. Existing approaches such as Reinforcement Learning from Human Feedback or Direct Preference Optimization typically require curated data and expensive optimization over billions of parameters, and eventually lead to persistent task-specific models. In this work, we introduce Preference alignment of Large Language Models via Residual Steering (PaLRS), a training-free method that exploits preference signals encoded in the residual streams of LLMs. From as few as one hundred preference pairs, PaLRS extracts lightweight, plug-and-play steering vectors that can be applied at inference time to push models toward preferred behaviors. We evaluate PaLRS on various small-to-medium-scale open-source LLMs, showing that PaLRS-aligned models achieve consistent gains on mathematical reasoning and code generation benchmarks while preserving baseline general-purpose performance. Moreover, when compared to DPO-aligned models, they perform better with huge time savings. Our findings highlight that PaLRS offers an effective, much more efficient and flexible alternative to standard preference optimization pipelines, offering a training-free, plug-and-play mechanism for alignment with minimal data.

Authorship Attribution in Multilingual Machine-Generated Texts

Lucio La Cava, Dominik Macko, Róbert Móro, Ivan Srba, Andrea Tagarelli

arXiv · 03 Aug 2025 · doi:10.48550/arXiv.2508.01656

As Large Language Models (LLMs) have reached human-like fluency and coherence, distinguishing machine-generated text (MGT) from human-written content becomes increasingly difficult. While early efforts in MGT detection have focused on binary classification, the growing landscape and diversity of LLMs require a more fine-grained yet challenging authorship attribution (AA), i.e., being able to identify the precise generator (LLM or human) behind a text. However, AA remains nowadays confined to a monolingual setting, with English being the most investigated one, overlooking the multilingual nature and usage of modern LLMs. In this work, we introduce the problem of Multilingual Authorship Attribution, which involves attributing texts to human or multiple LLM generators across diverse languages. Focusing on 18 languages – covering multiple families and writing scripts – and 8 generators (7 LLMs and the human-authored class), we investigate the multilingual suitability of monolingual AA methods, their cross-lingual transferability, and the impact of generators on attribution performance. Our results reveal that while certain monolingual AA methods can be adapted to multilingual settings, significant limitations and challenges remain, particularly in transferring across diverse language families, underscoring the complexity of multilingual AA and the need for more robust approaches to better match real-world scenarios.

Exploring LLMs’ Ability to Spontaneously and Conditionally Modify Moral Expressions through Text Manipulation

Candida Maria Greco, Lucio La Cava, Lorenzo Zangari, Andrea Tagarelli

Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 30 Jul 2025 · doi:10.18653/v1/2025.acl-long.883

Morality serves as the foundation of societal structure, guiding legal systems, shaping cultural values, and influencing individual self-perception. With the rise and pervasiveness of generative AI tools, and particularly Large Language Models (LLMs), concerns arise regarding how these tools capture and potentially alter moral dimensions through machine-generated text manipulation. Based on the Moral Foundation Theory, our work investigates this topic by analyzing the behavior of 12 LLMs among the most widely used Open and uncensored (i.e., ”abliterated”) models, and leveraging human-annotated datasets used in moral-related analysis. Results have shown varying levels of alteration of moral expressions depending on the type of text modification task and moral-related conditioning prompt.

Unveiling user dynamics in the evolving social debate on climate crisis during the conferences of the parties

Liliana Martirano, Lucio La Cava, Andrea Tagarelli

Pervasive and Mobile Computing · 30 May 2025 · doi:10.1016/j.pmcj.2025.102077

Social media have widely been recognized as a valuable proxy for investigating users’ opinions by echoing virtual venues where individuals engage in daily discussions on a wide range of topics. Among them, climate change is gaining momentum due to its large-scale impact, tangible consequences for society, and enduring nature. In this work, we investigate the social debate surrounding climate emergency, aiming to uncover the fundamental patterns that underlie the climate debate, thus providing valuable support for strategic and operational decision-making. To this purpose, we leverage Graph Mining and NLP techniques to analyze a large corpus of tweets spanning seven years pertaining to the Conference of the Parties (COP), the leading global forum for multilateral discussion on climate-related matters, based on our proposed framework, named NATMAC, which consists of three main modules designed to perform network analysis, topic modeling and affective computing tasks. Our contribution in this work is manifold - (i) we provide insights into the key social actors involved in the climate debate and their relationships, (ii) we unveil the main topics discussed during COPs within the social landscape, (iii) we assess the evolution of users’ sentiment and emotions across time, and (iv) we identify users’ communities based on multiple dimensions. Furthermore, our proposed approach exhibits the potential to scale up to other emergency issues, highlighting its versatility and potential for broader use in analyzing and understanding the increasingly debated emergent phenomena.

Safeguarding Decentralized Social Media: LLM Agents for Automating Community Rule Compliance

Lucio La Cava, Andrea Tagarelli

Online Social Networks and Media · 30 May 2025 · doi:10.1016/j.osnem.2025.100319

Ensuring content compliance with community guidelines is crucial for maintaining healthy online social environments. However, traditional human-based compliance checking struggles with scaling due to the increasing volume of user-generated content and a limited number of moderators. Recent advancements in Natural Language Understanding demonstrated by Large Language Models unlock new opportunities for automated content compliance verification. This work evaluates six AI-agents built on Open-LLMs for automated rule compliance checking in Decentralized Social Networks, a challenging environment due to heterogeneous community scopes and rules. By analyzing over 50,000 posts from hundreds of Mastodon servers, we find that AI-agents effectively detect non-compliant content, grasp linguistic subtleties, and adapt to diverse community contexts. Most agents also show high inter-rater reliability and consistency in score justification and suggestions for compliance. Human-based evaluation with domain experts confirmed the agents’ reliability and usefulness, rendering them promising tools for semi-automated or human-in-the-loop content moderation systems.

Visually Wired NFTs: Exploring the Role of Inspiration in Non-Fungible Tokens

Lucio La Cava, Davide Costa, Andrea Tagarelli

ACM Transactions on the Web · 24 May 2025 · doi:10.1145/3703411

The fervor for Non-Fungible Tokens (NFTs) attracted countless creators, leading to a Big Bang of digital assets driven by latent or explicit forms of inspiration, as in many creative processes. This work exploits Vision Transformers and graph-based modeling to delve into visual inspiration phenomena between NFTs over the years, i.e., the visual influence that can be detected whenever an NFT appears to be visually close to another that was published earlier in the market. Our goals include unveiling the main structural traits that shape visual inspiration networks, exploring the interrelation between visual inspiration and asset performances, investigating crypto influence on inspiration processes, and explaining the inspiration relationships among NFTs. Our findings unveil how the pervasiveness of inspiration led to a temporary saturation of the visual feature space, the impact of the dichotomy between inspiring and inspired NFTs on their financial performance, and an intrinsic self-regulatory mechanism between markets and inspiration waves. Our work can serve as a starting point for gaining a broader view of the evolution of Web3.

OpenTuringBench: An Open-Model-based Benchmark and Framework for Machine-Generated Text Detection and Attribution

Lucio La Cava, Andrea Tagarelli

arXiv · 15 Apr 2025 · doi:10.48550/arXiv.2504.11369

Open Large Language Models (OLLMs) are increasingly leveraged in generative AI applications, posing new challenges for detecting their outputs. We propose OpenTuringBench, a new benchmark based on OLLMs, designed to train and evaluate machine-generated text detectors on the Turing Test and Authorship Attribution problems. OpenTuringBench focuses on a representative set of OLLMs, and features a number of challenging evaluation tasks, including human/machine-manipulated texts, out-of-domain texts, and texts from previously unseen models. We also provide OTBDetector, a contrastive learning framework to detect and attribute OLLM-based machine-generated texts. Results highlight the relevance and varying degrees of difficulty of the OpenTuringBench tasks, with our detector achieving remarkable capabilities across the various tasks and outperforming most existing detectors. Resources are available on the OpenTuringBench Hugging Face repository.

Open Models, Closed Minds? On Agents Capabilities in Mimicking Human Personalities through Open Large Language Models

Lucio La Cava, Andrea Tagarelli

Proceedings of the AAAI Conference on Artificial Intelligence · 11 Apr 2025 · doi:10.1609/aaai.v39i2.32125

The emergence of unveiling human-like behaviors in Large Language Models (LLMs) has led to a closer connection between NLP and human psychology. However, research on the personalities exhibited by LLMs has largely been confined to limited investigations using individual psychological tests, primarily focusing on a small number of commercially licensed LLMs. This approach overlooks the extensive use and significant advancements observed in open-source LLMs. This work aims to address both the above limitations by conducting an in-depth investigation of a significant body of 12 LLM Agents based on the most representative Open models, through the two most well-known psychological assessment tests, namely Myers-Briggs Type Indicator (MBTI) and Big Five Inventory (BFI). Our approach involves evaluating the intrinsic personality traits of LLM agents and determining the extent to which these agents can mimic human personalities when conditioned by specific personalities and roles. Our findings unveil that (i) each LLM agent showcases distinct human personalities; (ii) personality-conditioned prompting produces varying effects on the agents, with only few successfully mirroring the imposed personality, while most of them being ``closed-minded’’ (i.e., they retain their intrinsic traits); and (iii) combining role and personality conditioning can enhance the agents’ ability to mimic human personalities. Our work represents a step up in understanding the dense relationship between NLP and human psychology through the lens of LLMs.

A survey on moral foundation theory and pre-trained language models: current advances and challenges

Lorenzo Zangari, Candida Maria Greco, Davide Picca, Andrea Tagarelli

AI & SOCIETY · 24 Mar 2025 · doi:10.1007/s00146-025-02225-w

Moral values have deep roots in early civilizations, codified within norms and laws that regulated societal order and the common good. They play a crucial role in understanding the psychological basis of human behavior and cultural orientation. The moral foundation theory (MFT) is a well-established framework that identifies the core moral foundations underlying the manner in which different cultures shape individual and social lives. Recent advancements in natural language processing, particularly pre-trained language models (PLMs), have enabled the extraction and analysis of moral dimensions from textual data. This survey presents a comprehensive review of MFT-informed PLMs, providing an analysis of moral tendencies in PLMs and their application in the context of MFT. We also review relevant datasets and lexicons and discuss trends, limitations, and future directions. By providing a structured overview of the intersection between PLMs and MFT, this work bridges moral psychology insights within the realm of PLMs, paving the way for further research and development in creating morally aware AI systems.

Heuristic-Informed Mixture of Experts for Link Prediction in Multilayer Networks

Lucio La Cava, Domenico Mandaglio, Lorenzo Zangari, Andrea Tagarelli

arXiv · 30 Jan 2025 · doi:10.48550/arXiv.2501.17557

Link prediction algorithms for multilayer networks are in principle required to effectively account for the entire layered structure while capturing the unique contexts offered by each layer. However, many existing approaches excel at predicting specific links in certain layers but struggle with others, as they fail to effectively leverage the diverse information encoded across different network layers. In this paper, we present MoE-ML-LP, the first Mixture-of-Experts (MoE) framework specifically designed for multilayer link prediction. Building on top of multilayer heuristics for link prediction, MoE-ML-LP synthesizes the decisions taken by diverse experts, resulting in significantly enhanced predictive capabilities. Our extensive experimental evaluation on real-world and synthetic networks demonstrates that MoE-ML-LP consistently outperforms several baselines and competing methods, achieving remarkable improvements of +60% in Mean Reciprocal Rank, +82% in Hits@1, +55% in Hits@5, and +41% in Hits@10. Furthermore, MoE-ML-LP features a modular architecture that enables the seamless integration of newly developed experts without necessitating the re-training of the entire framework, fostering efficiency and scalability to new experts, paving the way for future advancements in link prediction.

ME2-BERT: Are Events and Emotions what you need for Moral Foundation Prediction?

Lorenzo Zangari, Candida M. Greco, Davide Picca, Andrea Tagarelli

Proceedings of the 31st International Conference on Computational Linguistics · 25 Jan 2025 · url:https://aclanthology.org/2025.coling-main.638

Moralities, emotions, and events are complex aspects of human cognition, which are often treated separately since capturing their combined effects is challenging, especially due to the lack of annotated data. Leveraging their interrelations hence becomes crucial for advancing the understanding of human moral behaviors. In this work, we propose ME2-BERT, the first holistic framework for fine-tuning a pre-trained language model like BERT to the task of moral foundation prediction. ME2-BERT integrates events and emotions for learning domain-invariant morality-relevant text representations. Our extensive experiments show that ME2-BERT outperforms existing state-of-the-art methods for moral foundation prediction, with an average increase up to 35% in the out-of-domain scenario.

Talking the Talk Does Not Entail Walking the Walk: On the Limits of Large Language Models in Lexical Entailment Recognition

Candida Maria Greco, Lucio La Cava, Andrea Tagarelli

Findings of the Association for Computational Linguistics: EMNLP 2024 · 18 Nov 2024 · doi:10.18653/v1/2024.findings-emnlp.881

Verbs form the backbone of language, providing the structure and meaning to sentences. Yet, their intricate semantic nuances pose a longstanding challenge. Understanding verb relations through the concept of lexical entailment is crucial for comprehending sentence meanings and grasping verb dynamics. This work investigates the capabilities of eight Large Language Models in recognizing lexical entailment relations among verbs through differently devised prompting strategies and zero-/few-shot settings over verb pairs from two lexical databases, namely WordNet and HyperLex. Our findings unveil that the models can tackle the lexical entailment recognition task with moderately good performance, although at varying degree of effectiveness and under different conditions. Also, utilizing few-shot prompting can enhance the models’ performance. However, perfectly solving the task arises as an unmet challenge for all examined LLMs, which raises an emergence for further research developments on this topic.

A Survey on Moral Foundation Theory and Pre-Trained Language Models: Current Advances and Challenges

Lorenzo Zangari, Candida M. Greco, Davide Picca, Andrea Tagarelli

arXiv · 15 Oct 2024 · doi:10.48550/ARXIV.2409.13521

Moral values have deep roots in early civilizations, codified within norms and laws that regulated societal order and the common good. They play a crucial role in understanding the psychological basis of human behavior and cultural orientation. The Moral Foundation Theory (MFT) is a well-established framework that identifies the core moral foundations underlying the manner in which different cultures shape individual and social lives. Recent advancements in natural language processing, particularly Pre-trained Language Models (PLMs), have enabled the extraction and analysis of moral dimensions from textual data. This survey presents a comprehensive review of MFT-informed PLMs, providing an analysis of moral tendencies in PLMs and their application in the context of the MFT. We also review relevant datasets and lexicons and discuss trends, limitations, and future directions. By providing a structured overview of the intersection between PLMs and MFT, this work bridges moral psychology insights within the realm of PLMs, paving the way for further research and development in creating morally aware AI systems.

Is Contrasting All You Need? Contrastive Learning for the Detection and Attribution of AI-generated Text

Lucio La Cava, Davide Costa, Andrea Tagarelli

Frontiers in Artificial Intelligence and Applications · 12 Oct 2024 · doi:10.3233/FAIA240862

The significant progress in the development of Large Language Models has contributed to blurring the distinction between human and AI-generated text. The increasing pervasiveness of AI-generated text and the difficulty in detecting it poses new challenges for our society. In this paper, we tackle the problem of detecting and attributing AI-generated text by proposing WhosAI, a triplet-network contrastive learning framework designed to predict whether a given input text has been generated by humans or AI and to unveil the authorship of the text. Unlike most existing approaches, our proposed framework is conceived to learn semantic similarity representations from multiple generators at once, thus equally handling both detection and attribution tasks. Furthermore, WhosAI is model-agnostic and scalable to the release of new AI text-generation models by incorporating their generated instances into the embedding space learned by our framework. Experimental results on the TuringBench benchmark of 200K news articles show that our proposed framework achieves outstanding results in both the Turing Test and Authorship Attribution tasks, outperforming all the methods listed in the TuringBench benchmark leaderboards.

E2MoCase: A Dataset for Emotional, Event and Moral Observations in News Articles on High-impact Legal Cases

Candida M. Greco, Lorenzo Zangari, Davide Picca, Andrea Tagarelli

arXiv · 12 Oct 2024 · doi:10.48550/ARXIV.2409.09001

The way media reports on legal cases can significantly shape public opinion, often embedding subtle biases that influence societal views on justice and morality. Analyzing these biases requires a holistic approach that captures the emotional tone, moral framing, and specific events within the narratives. In this work we introduce E2MoCase, a novel dataset designed to facilitate the integrated analysis of emotions, moral values, and events within legal narratives and media coverage. By leveraging advanced models for emotion detection, moral value identification, and event extraction, E2MoCase offers a multi-dimensional perspective on how legal cases are portrayed in news articles.

Safeguarding Decentralized Social Media: LLM Agents for Automating Community Rule Compliance

Lucio La Cava, Andrea Tagarelli

arXiv · 12 Oct 2024 · doi:10.48550/ARXIV.2409.08963

Ensuring content compliance with community guidelines is crucial for maintaining healthy online social environments. However, traditional human-based compliance checking struggles with scaling due to the increasing volume of user-generated content and a limited number of moderators. Recent advancements in Natural Language Understanding demonstrated by Large Language Models unlock new opportunities for automated content compliance verification. This work evaluates six AI-agents built on Open-LLMs for automated rule compliance checking in Decentralized Social Networks, a challenging environment due to heterogeneous community scopes and rules. Analyzing over 50,000 posts from hundreds of Mastodon servers, we find that AI-agents effectively detect non-compliant content, grasp linguistic subtleties, and adapt to diverse community contexts. Most agents also show high inter-rater reliability and consistency in score justification and suggestions for compliance. Human-based evaluation with domain experts confirmed the agents’ reliability and usefulness, rendering them promising tools for semi-automated or human-in-the-loop content moderation systems.

Neural discovery of balance-aware polarized communities

Francesco Gullo, Domenico Mandaglio, Andrea Tagarelli

Machine Learning · 12 Aug 2024 · doi:10.1007/s10994-024-06581-4

Signed graphs are a model to depict friendly (positive) or antagonistic (negative) interactions (edges) among users (nodes). 2-Polarized-Communities (2pc) is a well-established combinatorial-optimization problem whose goal is to find two polarized communities from a signed graph, i.e., two subsets of nodes (disjoint, but not necessarily covering the entire node set) which exhibit a high number of both intra-community positive edges and negative inter-community edges. The state of the art in 2pc suffers from the limitations that (i) existing methods rely on a single (optimal) solution to a continuous relaxation of the problem in order to produce the ultimate discrete solution via rounding, and (ii) 2pc objective function comes with no control on size balance among communities. In this paper, we provide advances to the 2pc problem by addressing both these limitations, with a twofold contribution. First, we devise a novel neural approach that allows for soundly and elegantly explore a variety of suboptimal solutions to the relaxed 2pc problem, so as to pick the one that leads to the best discrete solution after rounding. Second, we introduce a generalization of 2pc objective function – termed -polarity – which fosters size balance among communities, and we incorporate it into the proposed machine-learning framework. Extensive experiments attest high accuracy of our approach, its superiority over the state of the art, and capability of function -polarity to discover high-quality size-balanced communities.

DyHANE: dynamic heterogeneous attributed network embedding through experience node replay

Liliana Martirano, Dino Ienco, Roberto Interdonato, Andrea Tagarelli

Applied Network Science · 15 Jul 2024 · doi:10.1007/s41109-024-00633-3

With real-world network systems typically comprising a large number of interactive components and inherently dynamic, Graph Continual Learning (GCL) has gained increasing popularity in recent years. Furthermore, most applications involve multiple entities and relationships with associated attributes, which has led to widely adopting Heterogeneous Information Networks (HINs) for capturing such rich structural and semantic meaning. In this context, we deal with the problem of learning multi-type node representations in a time evolving graph setting, harnessing the expressive power of Graph Neural Networks (GNNs). To this purpose, we propose a novel framework, named DyHANE—Dynamic Heterogeneous Attributed Network Embedding, which dynamically identifies a representative sample of multi-typed nodes as training set and updates the parameters of a GNN module, enabling the generation of up-to-date representations for all nodes in the network. We show the advantage of employing HINs on a data-incremental classification task. We compare the results obtained by DyHANE on a multi-step, incremental heterogeneous GAT model trained on a sample of changed and unchanged nodes, with the results obtained by either the same model trained from scratch or the same model trained solely on changed nodes. We demonstrate the effectiveness of the proposed approach in facing two major related challenges - (i) to avoid model re-train from scratch if only a subset of the network has been changed and (ii) to mitigate the risk of losing established patterns if the new nodes exhibit unseen properties. To the best of our knowledge, this is the first work that deals with the task of (deep) graph continual learning on HINs.

A meta-active learning approach exploiting instance importance

Sergio Flesca, Domenico Mandaglio, Francesco Scala, Andrea Tagarelli

Expert Systems with Applications · 08 Jun 2024 · doi:10.1016/j.eswa.2024.123320

Active learning is focused on minimizing the effort required to obtain labeled data by iteratively choosing fresh data samples for training a machine learning model. One of the primary challenges in active learning involves the selection of the most informative instances for labeling by an annotation oracle at each iteration. A viable approach is to develop an active learning strategy that aligns with the performance of a meta-learning model. This strategy evaluates the quality of previously selected instances and subsequently trains a machine learning model to predict the quality of instances to be labeled in the current iteration. This paper introduces a novel approach to learning for active learning, wherein instances are chosen for labeling based on their potential to induce the most substantial change in the current classifier. We explore various strategies for assessing the significance of an instance, taking into account variations in the learning gradient of the classification model. Our approach can be applied to any classifier that can be trained using gradient descent optimization. Here, we present a formulation that leverages a deep neural network model, which has not been extensively explored in existing learning-to-active-learn methodologies. Through experimental validation, our approach demonstrates promising results, especially in scenarios where there are limited initially labeled instances, particularly when the number of labeled instances per class is extremely limited.

Link Prediction on Multilayer Networks through Learning of Within-Layer and Across-Layer Node-Pair Structural Features and Node Embedding Similarity

Lorenzo Zangari, Domenico Mandaglio, Andrea Tagarelli

Proceedings of the ACM Web Conference 2024 · 17 May 2024 · doi:10.1145/3589334.3645646

Link prediction has traditionally been studied in the context of simple graphs, although real-world networks are inherently complex as they are often comprised of multiple interconnected components, or layers. Predicting links in such network systems, or multilayer networks, require to consider both the internal structure of a target layer as well as the structure of the other layers in a network, in addition to layer-specific node-attributes when available. This problem poses several challenges, even for graph neural network based approaches despite their successful and wide application to a variety of graph learning problems. In this work, we aim to fill a lack of multilayer graph representation learning methods designed for link prediction. Our proposal is a novel neural-network-based learning framework for link prediction on (attributed) multilayer networks, whose key idea is to combine (i) pairwise similarities of multilayer node embeddings learned by a graph neural network model, and (ii) structural features learned from both within-layer and across-layer link information based on overlapping multilayer neighborhoods. Extensive experimental results have shown that our framework consistently outperforms both single-layer and multilayer methods for link prediction on popular real-world multilayer networks, with an average percentage increase in AUC up to 38%. We make source code and evaluation data available at https://mlnteam-unical.github.io/resources/.

Polarization in Decentralized Online Social Networks

Lucio La Cava, Domenico Mandaglio, Andrea Tagarelli

ACM Web Science Conference · 04 May 2024 · doi:10.1145/3614419.3644013

Centralized social media platforms are currently experiencing a shift in user engagement, drawing attention to alternative paradigms like Decentralized Online Social Networks (DOSNs). The rising popularity of DOSNs finds its root in the accessibility of open-source software, enabling anyone to create a new instance (i.e., server) and participate in a decentralized network known as Fediverse. Despite this growing momentum, there has been a lack of studies addressing the effect of positive and negative interactions among instances within DOSNs. This work aims to fill this gap by presenting a preliminary examination of instances’ polarization in DOSNs, focusing on Mastodon — the most widely recognized decentralized social media platform, boasting over 10M users and nearly 20K instances to date. Our results suggest that polarization in the Fediverse emerges in unique ways, influenced by the desire to foster a federated environment between instances, also facilitating the isolation of instances that may pose potential risks to the Fediverse.

Evaluating GPT-3.5's Awareness and Summarization Abilities for European Constitutional Texts with Shared Topics

Candida M. Greco, A. Tagarelli

arXiv · 25 Jan 2024 · doi:10.48550/ARXIV.2401.14524

Constitutions are foundational legal documents that underpin the governmental and societal structures. As such, they are a reflection of a nation’s cultural and social uniqueness, but also contribute to establish topics of universal importance, like citizens’ rights and duties (RD). In this work, using the renowned GPT-3.5, we leverage generative large language models to understand constitutional passages that transcend national boundaries. A key contribution of our study is the introduction of a novel application of abstractive summarization on a multi-source collection of constitutional texts, with a focus on European countries’ constitution passages related to RD topics. Our results show the meaningfulness of GPT-3.5 to produce informative, coherent and faithful summaries capturing RD topics across European countries.

Open Models, Closed Minds? On Agents Capabilities in Mimicking Human Personalities through Open Large Language Models

Lucio La Cava, Andrea Tagarelli

arXiv · 13 Jan 2024 · doi:10.48550/arXiv.2401.07115

The emergence of unveiling human-like behaviors in Large Language Models (LLMs) has led to a closer connection between NLP and human psychology. Scholars have been studying the inherent personalities exhibited by LLMs and attempting to incorporate human traits and behaviors into them. However, these efforts have primarily focused on commercially-licensed LLMs, neglecting the widespread use and notable advancements seen in Open LLMs. This work aims to address this gap by employing a set of 12 LLM Agents based on the most representative Open models and subject them to a series of assessments concerning the Myers-Briggs Type Indicator (MBTI) test and the Big Five Inventory (BFI) test. Our approach involves evaluating the intrinsic personality traits of Open LLM agents and determining the extent to which these agents can mimic human personalities when conditioned by specific personalities and roles. Our findings unveil that (i) each Open LLM agent showcases distinct human personalities; (ii) personality-conditioned prompting produces varying effects on the agents, with only few successfully mirroring the imposed personality, while most of them being ``closed-minded’’ (i.e., they retain their intrinsic traits); and (iii) combining role and personality conditioning can enhance the agents’ ability to mimic human personalities. Our work represents a step up in understanding the dense relationship between NLP and human psychology through the lens of Open LLMs.

Drivers of social influence in the Twitter migration to Mastodon

Lucio La Cava, Luca Maria Aiello, Andrea Tagarelli

Scientific Reports · 07 Dec 2023 · doi:10.1038/s41598-023-48200-7

The migration of Twitter users to Mastodon following Elon Musk’s acquisition presents a unique opportunity to study collective behavior and gain insights into the drivers of coordinated behavior in online media. We analyzed the social network and the public conversations of about 75,000 migrated users and observed that the temporal trace of their migrations is compatible with a phenomenon of social influence, as described by a compartmental epidemic model of information diffusion. Drawing from prior research on behavioral change, we delved into the factors that account for variations of the effectiveness of the influence process across different Twitter communities. Communities in which the influence process unfolded more rapidly exhibit lower density of social connections, higher levels of signaled commitment to migrating, and more emphasis on shared identity and exchange of factual knowledge in the community discussion. These factors account collectively for 57% of the variance in the observed data. Our results highlight the joint importance of network structure, commitment, and psycho-linguistic aspects of social interactions in characterizing grassroots collective action, and contribute to deepen our understanding of the mechanisms that drive processes of behavior change of online groups.

Unraveling the NFT economy: A comprehensive collection of Non-Fungible Token transactions and metadata

Davide Costa, Lucio La Cava, Andrea Tagarelli

Data in Brief · 01 Dec 2023 · doi:10.1016/j.dib.2023.109749

Non-Fungible Tokens (NFTs) have emerged as the most representative application of blockchain technology in recent years, fostering the development of the Web3. Nonetheless, while the interest in NFTs rapidly boomed, creating unprecedented fervour in traders and creators, the demand for highly representative and up-to-date data to shed light on such an intriguing yet complex domain mostly remained unmet. To pursue this objective, we introduce a large collection of NFT transactions and associated metadata that correspond to trading operations between 2021 and 2023. Our developed dataset is the most extensive and representative in the NFT landscape to date, as it contains more than 70 M transactions performed by more than 6 M users across 36.3 M NFTs and 281 K collections. Moreover, this dataset boasts a wealth of metadata, including encoded textual descriptions and multimedia content, thus being suitable for a plethora of tasks relevant to database systems, AI, data science, Web and network science fields. This dataset represents a unique resource for researchers and industry practitioners to delve into the inner workings of NFTs through a multitude of perspectives, paving the way for unprecedented opportunities across multiple research fields.

GDPR Article Retrieval based on Domain-adaptive and Task-adaptive Legal Pre-trained Language Models

Andrea Simeri, Andrea Tagarelli

ACM Hypertext-2023 Workshop on 'Legal Information Retrieval meets Artificial Intelligence Workshop' (LIRAI) · 01 Dec 2023 ·

The General Data Protection Regulation (GDPR) is an European regulation on data protection and privacy for all individuals within the European Union (EU) and the European Economic Area (EEA), and for all foreign subjects dealing with European citizens data. Therefore, the GDPR has important legislation implications that hold beyond EU member states. In this paper, we address the problem of GDPR article retrieval through the use of pre-trained language models (PLMs). Our approach features several key aspects, which include both domain-general and domain-specific pre-trained BERT models, further powered by self-supervised task-adaptive pre-training stages, with or without data enrichment based on recitals. Our study endeavors to demonstrate the potential of PLMs in addressing the challenges posed by the GDPR’s intricate legal framework, thus ultimately facilitating efficient access to GDPR provisions for government agencies, law firms, legal professionals, and citizens alike.

Bringing order into the realm of Transformer-based language models for artificial intelligence and law

Candida M. Greco, Andrea Tagarelli

Artificial Intelligence and Law · 20 Nov 2023 · doi:10.1007/s10506-023-09374-7

Transformer-based language models (TLMs) have widely been recognized to be a cutting-edge technology for the successful development of deep-learning-based solutions to problems and applications that require natural language processing and understanding. Like for other textual domains, TLMs have indeed pushed the state-of-the-art of AI approaches for many tasks of interest in the legal domain. Despite the first Transformer model being proposed about six years ago, there has been a rapid progress of this technology at an unprecedented rate, whereby BERT and related models represent a major reference, also in the legal domain. This article provides the first systematic overview of TLM-based methods for AI-driven problems and tasks in the legal sphere. A major goal is to highlight research advances in this field so as to understand, on the one hand, how the Transformers have contributed to the success of AI in supporting legal processes, and on the other hand, what are the current limitations and opportunities for further research development.

Evolution of the Social Debate on Climate Crisis: Insights from Twitter During the Conferences of the Parties

Liliana Martirano, Lucio La Cava, Andrea Tagarelli

2023 International Conference on Information and Communication Technologies for Disaster Management (ICT-DM) · 13 Sep 2023 · doi:10.1109/ICT-DM58371.2023.10286927

Social media have long been recognized as a valuable proxy for investigating users’ opinions by echoing virtual venues where individuals engage in daily discussions on a wide range of topics. Among them, climate change is gaining momentum due to its large-scale impact, tangible consequences for society, and enduring nature. In this work, we investigate the social debate surrounding climate emergency, with particular emphasis on the Conference of the Parties (COP), the foremost global forum for multilateral discussion on climate-related matters. To this aim, we leverage graph mining and text mining techniques to analyze a large corpus of tweets spanning 7 years, aiming to uncover the fundamental patterns underlying the climate debate, thus providing valuable support for strategic and operational decision-making. Our contribution in this work is manifold - (i) we provide insights into the key social actors involved in the climate debate and their relationships, (ii) we unveil the main topics discussed during COPs within the social landscape, (iii) we assess the evolution of users’ sentiment and emotions across time. Furthermore, our proposed approach has the potential to scale up to other emergency issues, highlighting its versatility and potential for broader use in analyzing the increasingly debated emergent phenomena.

SONAR: Web-based Tool for Multimodal Exploration of Non-Fungible Token Inspiration Networks

Lucio La Cava, Davide Costa, Andrea Tagarelli

Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval · 18 Jul 2023 · doi:10.1145/3539618.3591821

In this work, we present SONAR, a web-based tool for multimodal exploration of Non-Fungible Token (NFT) inspiration networks. SONAR is conceived to support both creators and traders in the emerging Web3 by providing an interactive visualization of the inspiration-driven connections between NFTs, at both individual level and collection level. SONAR can hence be useful to identify new investment opportunities as well as anomalous inspirations. To demonstrate SONAR’s capabilities, we present an application to the largest and most representative dataset concerning the NFT landscape to date, showing how our proposed tool can scale and ensure high-level user experience up to millions of edges.

A combinatorial multi-armed bandit approach to correlation clustering

F. Gullo, D. Mandaglio, A. Tagarelli

Data Mining and Knowledge Discovery · 29 Jun 2023 · doi:10.1007/S10618-023-00937-5

Given a graph whose edges are assigned positive-type and negative-type weights, the problem of correlation clustering aims at grouping the graph vertices so as to minimize (resp. maximize) the sum of negative-type (resp. positive-type) intra-cluster weights plus the sum of positive-type (resp. negative-type) inter-cluster weights. In correlation clustering, it is typically assumed that the weights are readily available. This is a rather strong hypothesis, which is unrealistic in several scenarios. To overcome this limitation, in this work we focus on the setting where edge weights of a correlation-clustering instance are unknown, and they have to be estimated in multiple rounds, while performing the clustering. The clustering solutions produced in the various rounds provide a feedback to properly adjust the weight estimates, and the goal is to maximize the cumulative quality of the clusterings. We tackle this problem by resorting to the reinforcement-learning paradigm, and, specifically, we design for the first time a Combinatorial Multi-Armed Bandit (CMAB) framework for correlation clustering. We provide a variety of contributions, namely (1) formulations of the minimization and maximization variants of correlation clustering in a CMAB setting; (2) adaptation of well-established CMAB algorithms to the correlation-clustering context; (3) regret analyses to theoretically bound the accuracy of these algorithms; (4) design of further (heuristic) algorithms to have the probability constraint satisfied at every round (key condition to soundly adopt efficient yet effective algorithms for correlation clustering as CMAB oracles); (5) extensive experimental comparison among a variety of both CMAB and non-CMAB approaches for correlation clustering.

Show me your NFT and I tell you how it will perform: Multimodal representation learning for NFT selling price prediction

Davide Costa, Lucio La Cava, Andrea Tagarelli

Proceedings of the ACM Web Conference 2023 · 30 Apr 2023 · doi:10.1145/3543507.3583520

Non-Fungible Tokens (NFTs) represent deeds of ownership, based on blockchain technologies and smart contracts, of unique crypto assets on digital art forms (e.g., artworks or collectibles). In the spotlight after skyrocketing in 2021, NFTs have attracted the attention of crypto enthusiasts and investors intent on placing promising investments in this profitable market. However, the NFT financial performance prediction has not been widely explored to date. In this work, we address the above problem based on the hypothesis that NFT images and their textual descriptions are essential proxies to predict the NFT selling prices. To this purpose, we propose MERLIN, a novel multimodal deep learning framework designed to train Transformer-based language and visual models, along with graph neural network models, on collections of NFTs’ images and texts. A key aspect in MERLIN is its independence on financial features, as it exploits only the primary data a user interested in NFT trading would like to deal with, i.e., NFT images and textual descriptions. By learning dense representations of such data, a price-category classification task is performed by MERLIN models, which can also be tuned according to user preferences in the inference phase to mimic different risk-return investment profiles. Experimental evaluation on a publicly available dataset has shown that MERLIN models achieve significant performances according to several financial assessment criteria, fostering profitable investments, and also beating baseline machine-learning classifiers based on financial features.

Exploring domain and task adaptation of LamBERTa models for article retrieval on the Italian Civil Code

Andrea Simeri, Andrea Tagarelli

The 19th Conference on Information and Research science Connecting to Digital and Library science (IRCDL) · 21 Apr 2023 ·

This paper is concerned with AI-based NLP solutions to the law article retrieval problem, with application to the Italian legal domain and, particularly, to the Italian Civil Code. Based upon the current state-of- the-art on this topic, we revise our early LamBERTa framework in a twofold way relating its domain- adaptation feature - replacing the general-domain pre-trained model with a legal-specific one to fine-tune for the task of article retrieval, and delving into the injection of out-of-vocabulary legal terms into the models’ tokenizer. Extensive experimental evaluation based on different collections of query sets, along with qualitative analysis on the models’ prediction interpretability, have unveiled interesting findings about the combined effect of domain- and task-adaptation of an Italian BERT model on the task of law article retrieval.

Transformer-based language models for mental health issues: A survey

Candida M. Greco, Andrea Simeri, Andrea Tagarelli, Ester Zumpano

Pattern Recognition Letters · 01 Mar 2023 · doi:10.1016/j.patrec.2023.02.016

Early identification and prevention of mental health stresses and their outcomes has become of urgent importance worldwide. To this purpose, artificial intelligence provides a body of advanced computational tools that can effectively support decision-making clinical processes by modeling and analyzing the presence of a variety of mental health issues, particularly when these can be detected in text data. In this regard, Transformer-based language models (TLMs) have demonstrated exceptional efficacy in a number of NLP tasks also in the health domain. To the best of our knowledge, the use of TLMs for specifically addressing mental health issues has not been deeply investigated so far. In this paper, we aim to fill this gap in the literature by providing the first survey of methods using TLMs for text-based identification of mental health issues.

Drivers of social influence in the Twitter migration to Mastodon

Lucio La Cava, Luca Maria Aiello, Andrea Tagarelli

arXiv · 01 Jan 2023 · doi:10.48550/arXiv.2305.19056

The migration of Twitter users to Mastodon following Elon Musk’s acquisition presents a unique opportunity to study collective behavior and gain insights into the drivers of coordinated behavior in online media. We analyzed the social network and the public conversations of about 75,000 migrated users and observed that the temporal trace of their migrations is compatible with a phenomenon of social influence, as described by a compartmental epidemic model of information diffusion. Drawing from prior research on behavioral change, we delved into the factors that account for variations across different Twitter communities in the effectiveness of the spreading of the influence to migrate. Communities in which the influence process unfolded more rapidly exhibit lower density of social connections, higher levels of signaled commitment to migrating, and more emphasis on shared identity and exchange of factual knowledge in the community discussion. These factors account collectively for 57% of the variance in the observed data. Our results highlight the joint importance of network structure, commitment, and psycho-linguistic aspects of social interactions in describing grassroots collective action, and contribute to deepen our understanding of the mechanisms driving processes of behavior change of online groups.

Co-MLHAN: contrastive learning for multilayer heterogeneous attributed networks

Liliana Martirano, Lorenzo Zangari, Andrea Tagarelli

Applied Network Science · 20 Sep 2022 · doi:10.1007/s41109-022-00504-9

Graph representation learning has become a topic of great interest and many works focus on the generation of high-level, task-independent node embeddings for complex networks. However, the existing methods consider only few aspects of networks at a time. In this paper, we propose a novel framework, named Co-MLHAN, to learn node embeddings for networks that are simultaneously multilayer, heterogeneous and attributed. We leverage contrastive learning as a self-supervised and task-independent machine learning paradigm and define a cross-view mechanism between two views of the original graph which collaboratively supervise each other. We evaluate our framework on the entity classification task. Experimental results demonstrate the effectiveness of Co-MLHAN and its variant Co-MLHAN-SA, showing their capability of exploiting across-layer information in addition to other types of knowledge.

Learning to Active Learn by Gradient Variation based on Instance Importance

Sergio Flesca, Domenico Mandaglio, Francesco Scala, Andrea Tagarelli

2022 26th International Conference on Pattern Recognition (ICPR) · 21 Aug 2022 · doi:10.1109/ICPR56361.2022.9956039

A major challenge in active learning is to select the most informative instances to be labeled by an annotation oracle at each step. In this respect, one effective paradigm is to learn the active learning strategy that best suits the performance of a meta-learning model. This strategy first measures the quality of the instances selected in the previous steps and then trains a machine learning model that is used to predict the quality of instances to be labeled in the current step.In this paper, we propose a new approach of learning-to-active-learn that selects the instances to be labeled as the ones producing the maximum change to the current classifier. Our key idea is to select such instances according to their importance reflecting variations in the learning gradient of the classification model. Our approach can be instantiated with any classifier trainable via gradient descent optimization, and here we provide a formulation based on a deep neural network model, which has not deeply been investigated in existing learning-to-active-learn approaches. The experimental validation of our approach has shown promising results in scenarios characterized by relatively few initially labeled instances.

LawNet-Viz - A Web-based System to Visually Explore Networks of Law Article References

Lucio La Cava, Andrea Simeri, Andrea Tagarelli

Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval · 06 Jul 2022 · doi:10.1145/3477495.3531668

We present LawNet-Viz, a web-based tool for the modeling, analysis and visualization of law reference networks extracted from a statute law corpus. LawNet-Viz is designed to support legal research tasks and help legal professionals as well as laymen visually exploring the article connections built upon the explicit law references detected in the article contents. To demonstrate LawNet-Viz, we show its application to the Italian Civil Code (ICC), which exploits a recent BERT-based model fine-tuned on the ICC. LawNet-Viz is a system prototype that is planned for product development.

Information consumption and boundary spanning in Decentralized Online Social Networks: The case of Mastodon users

Lucio La Cava, Sergio Greco, Andrea Tagarelli

Online Social Networks and Media · 01 Jul 2022 · doi:10.1016/j.osnem.2022.100220

Decentralized Online Social Networks (DOSNs) represent a growing trend in the social media landscape, as opposed to the well-known centralized peers, which are often in the spotlight due to privacy concerns and a vision typically focused on monetization through user relationships. By exploiting open-source software, DOSNs allow users to create their own servers, or instances, thus favoring the proliferation of platforms that are independent yet interconnected with each other in a transparent way. Nonetheless, the resulting cooperation model, commonly known as the Fediverse, still represents a world to be fully discovered, since existing studies have mainly focused on a limited number of structural aspects of interest in DOSNs. In this work, we aim to fill a lack of study on user relations and roles in DOSNs, by taking two main actions - understanding the impact of decentralization on how users relate to each other within their membership instance and/or across different instances, and unveiling user roles that can explain two interrelated axes of social behavioral phenomena, namely information consumption and boundary spanning. To this purpose, we build our analysis on user networks from Mastodon, since it represents the most widely used DOSN platform. We believe that the findings drawn from our study on Mastodon users’ roles and information flow can pave a way for further development of fascinating research on DOSNs.

Network Analysis of the Information Consumption-Production Dichotomy in Mastodon User Behaviors

Lucio La Cava, Sergio Greco, Andrea Tagarelli

Proceedings of the International AAAI Conference on Web and Social Media · 31 May 2022 · doi:10.1609/icwsm.v16i1.19391

Decentralized Online Social Networks (DOSNs) are today an established alternative to the popular centralized counterparts. In this work, we push forward research on user behaviors in a decentralized context, by exploring the dichotomy between information consumption and production. Using the Mastodon user network as a proxy for the Fediverse landscape, we address two main research questions - Do the consumers, resp. producers, identified in one instance exhibit the same behavior consistently while interacting with other instances? and, Are there users who behave as consumers in one instance and simultaneously as producers in other instances, or vice versa? In this respect, our results reveal interesting traits of Mastodon users, yet unveil the emergence for further studies that can embrace other services in the Fediverse.

A Comparison of Transformer-Based Language Models on NLP Benchmarks

Candida Maria Greco, Andrea Tagarelli, Ester Zumpano

Lecture Notes in Computer Science · 01 Jan 2022 · doi:10.1007/978-3-031-08473-7_45

Since the advent of BERT, Transformer-based language models (TLMs) have shown outstanding effectiveness in several NLP tasks. In this paper, we aim at bringing order to the landscape of TLMs and their performance on important benchmarks for NLP. Our analysis sheds light on the advantages that some TLMs take over the others, but also unveils issues in making a complete and fair comparison in some situations.

When Correlation Clustering Meets Fairness Constraints

Francesco Gullo, Lucio La Cava, Domenico Mandaglio, Andrea Tagarelli

Lecture Notes in Computer Science · 01 Jan 2022 · doi:10.1007/978-3-031-18840-4_22

The study of fairness-related aspects in data analysis is an active field of research, which can be leveraged to understand and control specific types of bias in decision-making systems. A major problem in this context is fair clustering, i.e., grouping data objects that are similar according to a common feature space, while avoiding biasing the clusters against or towards particular types of classes or sensitive features. In this work, we focus on a correlation-clustering method we recently introduced, and experimentally assess its performance in a fairness-aware context. We compare it to state-of-the-art fair-clustering approaches, both in terms of classic clustering quality measures and fairness-related aspects. Experimental evidence on public real datasets has shown that our method yields solutions of higher quality than the competing methods according to classic clustering-validation criteria, without neglecting fairness aspects.

Graph convolutional and attention models for entity classification in multilayer networks

Lorenzo Zangari, Roberto Interdonato, Antonio Calió, Andrea Tagarelli

Applied Network Science · 08 Nov 2021 · doi:10.1007/s41109-021-00420-4

Graph Neural Networks (GNNs) are powerful tools that are nowadays reaching state of the art performances in a plethora of different tasks such as node classification, link prediction and graph classification. A challenging aspect in this context is to redefine basic deep learning operations, such as convolution, on graph-like structures, where nodes generally have unordered neighborhoods of varying size. State-of-the-art GNN approaches such as Graph Convolutional Networks (GCNs) and Graph Attention Networks (GATs) work on monoplex networks only, i.e., on networks modeling a single type of relation among an homogeneous set of nodes. The aim of this work is to generalize such approaches by proposing a GNN framework for representation learning and semi-supervised classification in multilayer networks with attributed entities, and arbitrary number of layers and intra-layer and inter-layer connections between nodes. We instantiate our framework with two new formulations of GAT and GCN models, namely ML-GCN and ML-GAT, specifically devised for general, attributed multilayer networks. The proposed approaches are evaluated on an entity classification task on nine widely used real-world network datasets coming from different domains and with different structural characteristics. Results show that both our proposed ML-GAT and ML-GCN methods provide effective and efficient solutions to the problem of entity classification in multilayer attributed networks, being faster to learn and offering better accuracy than the competitors. Furthermore, results show how our methods are able to take advantage of the presence of real attributes for the entities, in addition to arbitrary inter-layer connections between the nodes in the various layers.

Unsupervised law article mining based on deep pre-trained language representation models with application to the Italian civil code

Andrea Tagarelli, Andrea Simeri

Artificial Intelligence and Law · 15 Sep 2021 · doi:10.1007/s10506-021-09301-8

Modeling law search and retrieval as prediction problems has recently emerged as a predominant approach in law intelligence. Focusing on the law article retrieval task, we present a deep learning framework named LamBERTa, which is designed for civil-law codes, and specifically trained on the Italian civil code. To our knowledge, this is the first study proposing an advanced approach to law article prediction for the Italian legal system based on a BERT (Bidirectional Encoder Representations from Transformers) learning framework, which has recently attracted increased attention among deep learning approaches, showing outstanding effectiveness in several natural language processing and learning tasks. We define LamBERTa models by fine-tuning an Italian pre-trained BERT on the Italian civil code or its portions, for law article retrieval as a classification task. One key aspect of our LamBERTa framework is that we conceived it to address an extreme classification scenario, which is characterized by a high number of classes, the few-shot learning problem, and the lack of test query benchmarks for Italian legal prediction tasks. To solve such issues, we define different methods for the unsupervised labeling of the law articles, which can in principle be applied to any law article code system. We provide insights into the explainability and interpretability of our LamBERTa models, and we present an extensive experimental analysis over query sets of different type, for single-label as well as multi-label evaluation tasks. Empirical evidence has shown the effectiveness of LamBERTa, and also its superiority against widely used deep-learning text classifiers and a few-shot learner conceived for an attribute-aware prediction task.

Understanding the growth of the Fediverse through the lens of Mastodon

Lucio La Cava, Sergio Greco, Andrea Tagarelli

Applied Network Science · 01 Sep 2021 · doi:10.1007/s41109-021-00392-5

Open-source, Decentralized Online Social Networks (DOSNs) are emerging as alternatives to the popular yet centralized and profit-driven platforms like Facebook or Twitter. In DOSNs, users can set up their own server, or instance, while they can actually interact with users of other instances. Moreover, by adopting the same communication protocol, DOSNs become part of a massive social network, namely the Fediverse. Mastodon is the most relevant platform in the Fediverse to date, and also the one that has attracted attention from the research community. Existing studies are however limited to an analysis of a relatively outdated sample of Mastodon focusing on few aspects at a user level, while several open questions have not been answered yet, especially at the instance level. In this work, we aim at pushing forward our understanding of the Fediverse by leveraging the primary role of Mastodon therein. Our first contribution is the building of an up-to-date and highly representative dataset of Mastodon. Upon this new data, we have defined a network model over Mastodon instances and exploited it to investigate three major aspects - the structural features of the Mastodon network of instances from a macroscopic as well as a mesoscopic perspective, to unveil the distinguishing traits of the underlying federative mechanism; the backbone of the network, to discover the essential interrelations between the instances; and the growth of Mastodon, to understand how the shape of the instance network has evolved during the last few years, also when broading the scope to account for instances belonging to other platforms. Our extensive analysis of the above aspects has provided a number of findings that reveal distinguishing features of Mastodon and that can be used as a starting point for the discovery of all the DOSN Fediverse.

Community Detection in Multiplex Networks

Matteo Magnani, Obaida Hanteer, Roberto Interdonato, Luca Rossi, Andrea Tagarelli

ACM Computing Surveys · 08 May 2021 · doi:10.1145/3444688

A multiplex network models different modes of interaction among same-type entities. In this article, we provide a taxonomy of community detection algorithms in multiplex networks. We characterize the different algorithms based on various properties and we discuss the type of communities detected by each method. We then provide an extensive experimental evaluation of the reviewed methods to answer three main questions - to what extent the evaluated methods are able to detect ground-truth communities, to what extent different methods produce similar community structures, and to what extent the evaluated methods are scalable. One goal of this survey is to help scholars and practitioners to choose the right methods for the data and the task at hand, while also emphasizing when such choice is problematic.

Attribute based diversification of seeds for targeted influence maximization

Antonio Caliò, Andrea Tagarelli

Information Sciences · 01 Feb 2021 · doi:10.1016/J.INS.2020.08.093

Embedding diversity into knowledge discovery is important - the patterns mined will be more novel, more meaningful, and broader. Surprisingly, in the classic problem of influence maximization in social networks, relatively little study has been devoted to diversity and its integration into the objective function of an influence maximization method. In this work, we propose the integration of a categorical-based notion of seed diversity into the objective function of a targeted influence maximization problem. In this respect, we assume that the users of a social network are associated with a categorical dataset where each tuple expresses the profile of a user according to a predefined schema of categorical attributes. Upon this assumption, we design a class of monotone submodular functions specifically conceived for determining the diversity of the subset of categorical tuples associated with the seed users to be discovered. This allows us to develop an efficient approximate method, with a constant-factor guarantee of optimality. More precisely, we formulate the attribute-based diversity-sensitive targeted influence maximization problem under the state-of-the-art reverse influence sampling framework, and we develop a method, dubbed ADITUM, that ensures a (1-1/e-∊)-approximate solution under the general triggering diffusion model. Extensive experimental evaluation based on real-world networks as well as synthetically generated data has shown the meaningfulness and uniqueness of our proposed class of set diversity functions and of the ADITUM algorithm, also in comparison with methods that exploit numerical-attribute-based diversity and topology-driven diversity in influence maximization.

Correlation Clustering with Global Weight Bounds

Domenico Mandaglio, Andrea Tagarelli, Francesco Gullo

Lecture Notes in Computer Science · 01 Jan 2021 · doi:10.1007/978-3-030-86520-7_31

Given a set of objects and nonnegative real weights expressing “positive” and “negative” feeling of clustering any two objects together, min-disagreement correlation clustering partitions the input object set so as to minimize the sum of the intra-cluster negative-type weights plus the sum of the inter-cluster positive-type weights. Min-disagreement correlation clustering is -hard, but efficient constant-factor approximation algorithms exist if the weights are bounded in some way. The weight bounds so far studied in the related literature are mostly local, as they are required to hold for every object-pair. In this paper, we introduce the problem of min-disagreement correlation clustering with global weight bounds, i.e., constraints to be satisfied by the input weights altogether. Our main result is a sufficient condition that establishes when any algorithm achieving a certain approximation under the probability constraint keeps the same guarantee on an input that violates the constraint. This extends the range of applicability of the most prominent existing correlation-clustering algorithms, including the popular Pivot, thus providing benefits, both theoretical and practical. Experiments demonstrate the usefulness of our approach, in terms of both worthiness of employing existing efficient algorithms, and guidance on the definition of weights from feature vectors in a task of fair clustering.

Research

All — Selected Publications, since 2021

2026

2025

2024

2023

2022

2021