Papers by Marco Guerini
Countering Hateful and Offensive Speech Online - Open Challenges (2024.emnlp-tutorials)
Copied to clipboard
Leon Derczynski, Marco Guerini, Debora Nozza, Flor Miriam Plaza-del-Arco, Jeffrey Sorensen, Marcos Zampieri
| Challenge: | a comprehensive understanding of the field is needed to maintain respectful and inclusive online environments. |
| Approach: | This tutorial aims to provide attendees with a comprehensive understanding of the field by delving into essential dimensions such as multilingualism, counter-narrative generation, a hands-on session with one of the most popular APIs for detecting hate speech, fairness, and ethics in AI, and the use of recent advanced approaches. |
| Outcome: | This tutorial aims to provide attendees with a comprehensive understanding of the field by delving into essential dimensions such as multilingualism, counter-narrative generation, a hands-on session with one of the most popular APIs for detecting hate speech, fairness, and ethics in AI, and the use of recent advanced approaches. |
Do LLMs suffer from Multi-Party Hangover? A Diagnostic Approach to Addressee Recognition and Response Selection in Conversations (2024.emnlp-main)
Copied to clipboard
| Challenge: | Conventional evaluation methods often overlook variances in model behavior across different levels of structural complexity on interaction graphs. |
| Approach: | They propose a methodological pipeline to investigate model performance across structural attributes of conversations. |
| Outcome: | The proposed method analyzes the performance of an LLM to classify multi-party conversations . it shows that response selection relies more on the textual content of conversations compared to addressee recognition . |
CrisiText: A dataset of warning messages for LLM training in emergency communication (2026.findings-eacl)
Copied to clipboard
| Challenge: | Identifying threats and mitigating their potential damage during crisis situations is paramount for safeguarding endangered individuals. |
| Approach: | They present a large-scale dataset for the generation of warning messages across 13 different types of crisis scenarios. |
| Outcome: | The proposed dataset contains more than 400,000 warning messages (spanning almost 18,000 crisis situations) aimed at assisting civilians during and after such events. |
Countering Misinformation via Emotional Response Generation (2023.emnlp-main)
Copied to clipboard
| Challenge: | Social media platforms (SMPs) are one of the most effective ways to spread misinformation by engaging in constructive dialogue with users who spread – often in good faith – misleading messages. |
| Approach: | They propose to use social correction to engage in constructive dialogue with users who spread misleading messages. |
| Outcome: | The proposed dataset shows that it improves on previous studies on claim-response pairs and the author-reviewer pipeline. |
Agreeing to Disagree: Annotating Offensive Language Datasets with Annotators’ Disagreement (2021.emnlp-main)
Copied to clipboard
| Challenge: | supervised learning is a key component of offensive language detection, but there is little attention given to the quality of annotated data. |
| Approach: | They propose to examine the level of agreement among annotators while selecting data to create offensive language datasets, a task involving a high level of subjectivity. |
| Outcome: | The proposed datasets show that annotators' agreement has a strong effect on classifiers performance and robustness. |
Putting Context in Context: the Impact of Discussion Structure on Text Classification (2024.eacl-long)
Copied to clipboard
| Challenge: | Current text classification approaches focus on the content to be classified, but contextual information is neglected in many cases. |
| Approach: | They propose to integrate contextual information into a transformer-based model by feeding it as natural language input into . they also experiment with different amounts of training data and analyse local discussion networks in a privacy-compliant way. |
| Outcome: | The proposed model can be generalized to other datasets and is privacy-compliant. |
FASTDial: Abstracting Dialogue Policies for Fast Development of Task Oriented Agents (P19-3)
Copied to clipboard
| Challenge: | Existing tools for creating task oriented dialogue agents are very flexible and require domain expertise to design them. |
| Approach: | They propose a framework for task oriented dialogue agents built on top of the OpenDial toolkit. |
| Outcome: | The proposed framework reduces programming effort and domain expert training time by hiding many implementation details. |
Fine-tuning with HED-IT: The impact of human post-editing for dialogical language models (2024.findings-acl)
Copied to clipboard
Daniela Occhipinti, Michele Marchi, Irene Mondella, Huiyuan Lai, Felice Dell’Orletta, Malvina Nissim, Marco Guerini
| Challenge: | a recent study has focused on the quality of data generated by automatic methods for fine-tuning Language Models in languages less resourced than English. |
| Approach: | They investigate whether human intervention improves the quality of machine-generated dialogues . they use a large-scale dataset to fine-tune three different sizes of an LM . |
| Outcome: | The results show that human intervention can improve the quality of training data . larger models are less sensitive to data quality, while smaller models are more sensitive . |
Towards Knowledge-Grounded Counter Narrative Generation for Hate Speech (2021.findings-acl)
Copied to clipboard
| Challenge: | Existing approaches to combat online hatred using informed textual responses - called counter narratives - produce generic/repetitive responses and lack grounded and up-to-date evidence such as facts, statistics, or examples. |
| Approach: | They propose to automatically generate counter narratives using an external knowledge repository to provide more informative content to fight online hatred. |
| Outcome: | The proposed pipeline can generate suitable and informative counter narratives in in-domain and cross-domain settings. |
Human-Machine Collaboration Approaches to Build a Dialogue Dataset for Hate Speech Countering (2022.emnlp-main)
Copied to clipboard
| Challenge: | a new approach to combat online hate speech is being proposed for NLG . existing methods to train NLG are limited to 2-turn interactions, while in real life, interactions can consist of multiple turns. |
| Approach: | They propose to combine human annotators with machine generated dialogues to create a dataset . DIALOCONAN is the first dataset comprising over 3000 fictitious multi-turn dialogues . |
| Outcome: | The proposed approach combines human experts over machine generated dialogues . it is the first dataset comprising over 3000 fictitious multi-turn dialogues between a hater and an NGO operator . |
Generating Counter Narratives against Online Hate Speech: Data and Strategies (2020.acl-main)
Copied to clipboard
| Challenge: | Hate Speech (HS) is a pervasive issue that spreads quickly and widely . research has focused on avoiding undesired effects that come with content moderation . |
| Approach: | They propose to use large scale unsupervised language models to generate responses to hate effectively using large scale models. |
| Outcome: | The proposed methods lack quality data and produce generic/repetitive responses. |
Regrexit or not Regrexit: Aspect-based Sentiment Analysis in Polarized Contexts (2020.coling-main)
Copied to clipboard
| Challenge: | Aspect-based Sentiment Analysis (ABSA) aims at capturing sentiment expressed toward each aspect of a target entity. |
| Approach: | They propose to extend the task of Aspect-based Sentiment Analysis (ABSA) toward affect and emotion representation in polarized settings. |
| Outcome: | The proposed model captures aspect-based polarization from newspapers regarding the Brexit scenario of 1.2m entities at sentence-level. |
NLP for Counterspeech against Hate and Misinformation (CSHAM) (2025.acl-tutorials)
Copied to clipboard
| Challenge: | tutorial aims to show how counterspeech is used to tackle abuse and misinformation by individuals, activists and organisations. |
| Approach: | tutorial aims to show how counterspeech is currently used to tackle abuse and misinformation . will also show how Natural Language Processing (NLP) and Generation (NLG) can be applied to automate its production. |
| Outcome: | The tutorial will bring diverse multidisciplinary perspectives to safety research . case studies from industry and public policy will be included . |
When Harry Meets Superman: The Role of The Interlocutor in Persona-Based Dialogue Generation (2025.acl-long)
Copied to clipboard
| Challenge: | In recent years, large language models (LLMs) have proven effective in generating coherent and contextually appropriate responses. |
| Approach: | They examine the ability of a model to adapt to the interlocutor's profile by masking or disclosing information about interlucutor . |
| Outcome: | The proposed model generalises well across topics, but struggles with unfamiliar interlocutors. |
Human-in-the-Loop for Data Collection: a Multi-Target Counter Narrative Dataset to Fight Online Hate Speech (2021.acl-long)
Copied to clipboard
| Challenge: | Existing studies on generating hate speech/counter narratives have failed to reach high-quality datasets. |
| Approach: | They propose a human-in-the-loop data collection methodology that refines a generative language model iteratively by using its own data from previous loops to generate new training samples. |
| Outcome: | The proposed method is the only expert-based multi-target HS/CN dataset available to the community. |
Using Pre-Trained Language Models for Producing Counter Narratives Against Hate Speech: a Comparative Study (2022.findings-acl)
Copied to clipboard
| Challenge: | Autoregressive models combined with stochastic decodings are the most promising for generating CNs with regard to an unseen target of hate. |
| Approach: | They propose to use pre-trained language models to generate counter-narratives in English by adding an automatic post-editing step to refine generated CNs. |
| Outcome: | The proposed pipeline could be used to generate counter-narratives in English using pre-trained language models and stochastic decoding mechanisms. |
Toward Stance-based Personas for Opinionated Dialogues (2020.findings-emnlp)
Copied to clipboard
| Challenge: | chit-chat neural models lacking specificity and coherence, argues a new study on stance-based personas . stancebased personal representations lack generalization capability, allowing agents to sustain personal points of view both within the same conversation and across different discussions. |
| Approach: | They propose to investigate stance-based persona representations and their impact on claim generation by using a conversational dataset. |
| Outcome: | The proposed dataset shows that stance-based personas grasp abstract and profound aspects of the author persona. |
Basque and Spanish Counter Narrative Generation: Data Creation and Evaluation (2024.lrec-main)
Copied to clipboard
| Challenge: | Davidson et al.: hate speech is a growing media presence, but research on generating CNs has been limited . he says a new dataset for CN generation is available for basque and spanish . this dataset is based on a multilingual encoder-decoder model . |
| Approach: | They propose a new Basque and Spanish dataset for automatic CN generation . they use machine translation and professional post-edition to generate CNs in both languages . |
| Outcome: | The proposed datasets show that training on post-edited data improves generation over monolingual settings . similar results in zero-shot crosslingual evaluations show multilingual data augmentation outperforms training in English and Spanish . |
CONAN - COunter NArratives through Nichesourcing: a Multilingual Dataset of Responses to Fight Online Hate Speech (P19-1)
Copied to clipboard
| Challenge: | Davidson et al., 2017): social media platforms and governmental organizations have taken steps to tackle hate speech . Davidson and Norton, 2017: a dataset of hate-speech/counter-narrative pairs is created . authors: identifying hate speech is challenging for the broadness and nuances in cultures and languages . |
| Approach: | They propose to build a large-scale, multilingual, expert-based dataset of hate-speech/counter-narrative pairs . they provide additional annotations about expert demographics, hate and response type . |
| Outcome: | The proposed dataset provides an analysis of hate-speech/counter-narrative pairs in three languages. |
PRODIGy: a PROfile-based DIalogue Generation dataset (2024.findings-naacl)
Copied to clipboard
| Challenge: | Existing profiles-based dialogue datasets lack explicit profile representations or are difficult to collect. |
| Approach: | They propose a dataset that brings together multiple profiles for each speaker, and then integrates them together to provide a more comprehensive profile dimension set for generative language models. |
| Outcome: | The PRODIGy dataset provides a more comprehensive profile dimension set for each speaker. |
Is Safer Better? The Impact of Guardrails on the Argumentative Strength of LLMs in Hate Speech Countering (2024.emnlp-main)
Copied to clipboard
| Challenge: | Automated responses lack argumentative richness which characterises expert-produced counterspeech. |
| Approach: | They propose to automate counterspeech generation by investigating tension between helpfulness and harmlessness of LLMs and to assess whether presence of safety guardrails hinders quality of generations. |
| Outcome: | The proposed approach produces more cogent responses that lack argumentative richness which characterises expert-produced counterspeech. |
EuroVerdict: A Multilingual Dataset for Verdict Generation Against Misinformation (2025.findings-acl)
Copied to clipboard
| Challenge: | a global issue that shapes public discourse shapes opinion and decision-making . many multilingual work has focused on claim verification rather than generating explanatory verdicts . |
| Approach: | They propose a multilingual dataset designed for verdict generation covering eight European languages. |
| Outcome: | The EuroVerdict dataset covers claims, manual verdicts, and supporting evidence . it is compared with other datasets in eight European languages . |
NLP for Counterspeech against Hate: A Survey and How-To Guide (2024.findings-naacl)
Copied to clipboard
| Challenge: | Recent studies have focused on the challenges of analysing, collecting, classifying, and automatically generating counterspeech, to reduce the huge burden of manually producing it. |
| Approach: | They propose a guide for doing research on counterspeech, with detailed examples and best practices that can be learnt from the NLP community. |
| Outcome: | The proposed strategies can reduce online and offline violence while preserving the freedom of speech of the users. |
Benchmarking the Generation of Fact Checking Explanations (2023.tacl-1)
Copied to clipboard
| Challenge: | Automating fact-checking is a time-consuming task that cannot keep up with the ever-increasing amount of fake news produced daily. |
| Approach: | They propose to automate the process of fact-checking by generating justifications from textual explanations of why a claim is classified as either true or false. |
| Outcome: | The proposed approach improves summarization performance over unstructured knowledge and with two datasets with different styles and structures. |
First-AID: the first Annotation Interface for grounded Dialogues (2025.acl-demo)
Copied to clipboard
| Challenge: | Existing tools to fine-tune Large Language Models for specific tasks are limited due to financial constraints and limited availability of human experts. |
| Approach: | They propose a human-in-the-loop framework for the knowledge-driven generation of synthetic dialogues using LLM prompting that implements different strategies of data collection that require different user intervention during dialogue generation. |
| Outcome: | The proposed framework reduces post-editing efforts and improves quality of generated dialogues. |