The potential effectiveness of counterspeech as a hate speech mitigation strategy is attracting increasing interest in the NLG research community, particularly towards the task of automatically producing it. However, automatically generated responses often lack the argumentative richness which characterises expert-produced counterspeech. In this work, we focus on two aspects of counterspeech generation to produce more cogent responses. First, by investigating the tension between helpfulness and harmlessness of LLMs, we test whether the presence of safety guardrails hinders the quality of the generations. Secondly, we assess whether attacking a specific component of the hate speech results in a more effective argumentative strategy to fight online hate. By conducting an extensive human and automatic evaluation, we show how the presence of safety guardrails can be detrimental also to a task that inherently aims at fostering positive social interactions. Moreover, our results show that attacking a specific component of the hate speech, and in particular its implicit negative stereotype and its hateful parts, leads to higher-quality generations.

Assessing the performance of systems to classify Multi-Party Conversations (MPC) is challenging due to the interconnection between linguistic and structural characteristics of conversations. Conventional evaluation methods often overlook variances in model behavior across different levels of structural complexity on interaction graphs. In this work, we propose a methodological pipeline to investigate model performance across specific structural attributes of conversations. As a proof of concept we focus on Response Selection and Addressee Recognition tasks, to diagnose model weaknesses. To this end, we extract representative diagnostic subdatasets with a fixed number of users and a good structural variety from a large and open corpus of online MPCs. We further frame our work in terms of data minimization, avoiding the use of original usernames to preserve privacy, and propose alternatives to using original text messages. Results show that response selection relies more on the textual content of conversations, while addressee recognition requires capturing their structural dimension. Using an LLM in a zero-shot setting, we further highlight how sensitivity to prompt variations is task-dependent.

In today’s digital age, hate speech and offensive speech online pose a significant challenge to maintaining respectful and inclusive online environments. 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. In addition, the tutorial aims to foster collaboration and inspire participants to create safer online spaces by detecting and mitigating hate speech.

In recent years, counterspeech has emerged as one of the most promising strategies to fight online hate. These non-escalatory responses tackle online abuse while preserving the freedom of speech of the users, and can have a tangible impact in reducing online and offline violence. Recently, there has been growing interest from the Natural Language Processing (NLP) community in addressing the challenges of analysing, collecting, classifying, and automatically generating counterspeech, to reduce the huge burden of manually producing it. In particular, researchers have taken different directions in addressing these challenges, thus providing a variety of related tasks and resources. In this paper, we provide a guide for doing research on counterspeech, by describing - with detailed examples - the steps to undertake, and providing best practices that can be learnt from the NLP studies on this topic. Finally, we discuss open challenges and future directions of counterspeech research in NLP.

Providing dialogue agents with a profile representation can improve their consistency and coherence, leading to better conversations. However, current profile-based dialogue datasets for training such agents contain either explicit profile representations that are simple and dialogue-specific, or implicit representations that are difficult to collect. In this work, we introduce the PRODIGy (PROfile-based DIalogue Generation) dataset, which brings diverse representations together, providing a more comprehensive profile dimension set for each speaker. This resource comprises more than 20k dialogues, sourced from movie scripts, aligned with speaker representations such as communication style, biography, personality and gender. Initial experiments with diverse baselines show that providing generative language models with these aspects of a profile, both separately and jointly, enhances models’ performance. This improvement holds true in both in-domain and cross-domain settings, for both fine-tuned and instruction-based LLMs.

Automatic methods for generating and gathering linguistic data have proven effective for fine-tuning Language Models (LMs) in languages less resourced than English. Still, while there has been emphasis on data quantity, less attention has been given to its quality. In this work, we investigate the impact of human intervention on machine-generated data when fine-tuning dialogical models. In particular, we study (1) whether post-edited dialogues exhibit higher perceived quality compared to the originals that were automatically generated; (2) whether fine-tuning with post-edited dialogues results in noticeable differences in the generated outputs; and (3) whether post-edited dialogues influence the outcomes when considering the parameter size of the LMs. To this end we created HED-IT, a large-scale dataset where machine-generated dialogues are paired with the version post-edited by humans. Using both the edited and unedited portions of HED-IT, we fine-tuned three different sizes of an LM. Results from both human and automatic evaluation show that the different quality of training data is clearly perceived and it has an impact also on the models trained on such data. Additionally, our findings indicate that larger models are less sensitive to data quality, whereas this has a crucial impact on smaller models. These results enhance our comprehension of the impact of human intervention on training data in the development of high-quality LMs.

Current text classification approaches usually focus on the content to be classified. Contextual aspects (both linguistic and extra-linguistic) are usually neglected, even in tasks based on online discussions. Still in many cases the multi-party and multi-turn nature of the context from which these elements are selected can be fruitfully exploited. In this work, we propose a series of experiments on a large dataset for stance detection in English, in which we evaluate the contribution of different types of contextual information, i.e. linguistic, structural and temporal, by feeding them as natural language input into a transformer-based model. We also experiment with different amounts of training data and analyse the topology of local discussion networks in a privacy-compliant way. Results show that structural information can be highly beneficial to text classification but only under certain circumstances (e.g. depending on the amount of training data and on discussion chain complexity). Indeed, we show that contextual information on smaller datasets from other classification tasks does not yield significant improvements. Our framework, based on local discussion networks, allows the integration of structural information while minimising user profiling, thus preserving their privacy.

Counter Narratives (CNs) are non-negative textual responses to Hate Speech (HS) aiming at defusing online hatred and mitigating its spreading across media. Despite the recent increase in HS content posted online, research on automatic CN generation has been relatively scarce and predominantly focused on English. In this paper, we present CONAN-EUS, a new Basque and Spanish dataset for CN generation developed by means of Machine Translation (MT) and professional post-edition. Being a parallel corpus, also with respect to the original English CONAN, it allows to perform novel research on multilingual and crosslingual automatic generation of CNs. Our experiments on CN generation with mT5, a multilingual encoder-decoder model, shows that generation greatly benefits from training on post-edited data, as opposed to relying on silver MT data only. These results are confirmed by their correlation with a qualitative manual evaluation, demonstrating that manually revised training data remains crucial for the quality of the generated CNs. Furthermore, multilingual data augmentation improves results over monolingual settings for structurally similar languages such as English and Spanish, while being detrimental for Basque, a language isolate. Similar findings occur in zero-shot crosslingual evaluations, where model transfer (fine-tuning in English and generating in a different target language) outperforms fine-tuning mT5 on machine translated data for Spanish but not for Basque. This provides an interesting insight into the asymmetry in the multilinguality of generative models, a challenging topic which is still open to research. Data and code will be made publicly available upon publication.

The proliferation of misinformation on social media platforms (SMPs) poses a significant danger to public health, social cohesion and ultimately democracy. Previous research has shown how social correction can be an effective way to curb misinformation, by engaging directly in a constructive dialogue with users who spread – often in good faith – misleading messages. Although professional fact-checkers are crucial to debunking viral claims, they usually do not engage in conversations on social media. Thereby, significant effort has been made to automate the use of fact-checker material in social correction; however, no previous work has tried to integrate it with the style and pragmatics that are commonly employed in social media communication. To fill this gap, we present VerMouth, the first large-scale dataset comprising roughly 12 thousand claim-response pairs (linked to debunking articles), accounting for both SMP-style and basic emotions, two factors which have a significant role in misinformation credibility and spreading. To collect this dataset we used a technique based on an author-reviewer pipeline, which efficiently combines LLMs and human annotators to obtain high-quality data. We also provide comprehensive experiments showing how models trained on our proposed dataset have significant improvements in terms of output quality and generalization capabilities.

Recent computational approaches for combating online hate speech involve the automatic generation of counter narratives by adapting Pretrained Transformer-based Language Models (PLMs) with human-curated data. This process, however, can produce in-domain overfitting, resulting in models generating acceptable narratives only for hatred similar to training data, with little portability to other targets or to real-world toxic language. This paper introduces novel attention regularization methodologies to improve the generalization capabilities of PLMs for counter narratives generation. Overfitting to training-specific terms is then discouraged, resulting in more diverse and richer narratives. We experiment with two attention-based regularization techniques on a benchmark English dataset. Regularized models produce better counter narratives than state-of-the-art approaches in most cases, both in terms of automatic metrics and human evaluation, especially when hateful targets are not present in the training data. This work paves the way for better and more flexible counter-speech generation models, a task for which datasets are highly challenging to produce.

Hate speech is a critical problem in our society and social media platforms are often an amplifier for this phenomenon. Recently the use of Counter Narratives (informative and non-aggressive responses) has been proposed as a viable solution to counter hateful content that goes beyond simple detection-removal strategies. In this paper we present a novel approach along this line of research, which utilizes the implied statement (bias) expressed in the hate speech to retrieve an appropriate counter narrative. To this end, we first trained and tested several LMs that, given a hateful post, generate the underlying bias and the target group. Then, for the counter narrative selection task, we experimented with several methodologies that either use or not use the implied bias during the process. Experiments show that using the target group information allows the system to better focus on relevant content and that implied statement for selecting counter narratives is better than the corresponding standard approach that does not use it. To our knowledge, this is the first attempt to build an automatic selection tool that uses hate speech implied bias to drive Counter Narrative selection.

Fighting misinformation is a challenging, yet crucial, task. Despite the growing number of experts being involved in manual fact-checking, this activity is time-consuming and cannot keep up with the ever-increasing amount of fake news produced daily. Hence, automating this process is necessary to help curb misinformation. Thus far, researchers have mainly focused on claim veracity classification. In this paper, instead, we address the generation of justifications (textual explanation of why a claim is classified as either true or false) and benchmark it with novel datasets and advanced baselines. In particular, we focus on summarization approaches over unstructured knowledge (i.e., news articles) and we experiment with several extractive and abstractive strategies. We employed two datasets with different styles and structures, in order to assess the generalizability of our findings. Results show that in justification production summarization benefits from the claim information, and, in particular, that a claim-driven extractive step improves abstractive summarization performances. Finally, we show that although cross-dataset experiments suffer from performance degradation, a unique model trained on a combination of the two datasets is able to retain style information in an efficient manner.

Fighting online hate speech is a challenge that is usually addressed using Natural Language Processing via automatic detection and removal of hate content. Besides this approach, counter narratives have emerged as an effective tool employed by NGOs to respond to online hate on social media platforms. For this reason, Natural Language Generation is currently being studied as a way to automatize counter narrative writing. However, the existing resources necessary to train NLG models are limited to 2-turn interactions (a hate speech and a counter narrative as response), while in real life, interactions can consist of multiple turns. In this paper, we present a hybrid approach for dialogical data collection, which combines the intervention of human expert annotators over machine generated dialogues obtained using 19 different configurations. The result of this work is DIALOCONAN, the first dataset comprising over 3000 fictitious multi-turn dialogues between a hater and an NGO operator, covering 6 targets of hate.

In this work, we present an extensive study on the use of pre-trained language models for the task of automatic Counter Narrative (CN) generation to fight online hate speech in English. We first present a comparative study to determine whether there is a particular Language Model (or class of LMs) and a particular decoding mechanism that are the most appropriate to generate CNs. Findings show that autoregressive models combined with stochastic decodings are the most promising. We then investigate how an LM performs in generating a CN with regard to an unseen target of hate. We find out that a key element for successful ‘out of target’ experiments is not an overall similarity with the training data but the presence of a specific subset of training data, i. e. a target that shares some commonalities with the test target that can be defined a-priori. We finally introduce the idea of a pipeline based on the addition of an automatic post-editing step to refine generated CNs.

Undermining the impact of hateful content with informed and non-aggressive responses, called counter narratives, has emerged as a possible solution for having healthier online communities. Thus, some NLP studies have started addressing the task of counter narrative generation. Although such studies have made an effort to build hate speech / counter narrative (HS/CN) datasets for neural generation, they fall short in reaching either high-quality and/or high-quantity. In this paper, we propose a novel human-in-the-loop data collection methodology in which a generative language model is refined iteratively by using its own data from the previous loops to generate new training samples that experts review and/or post-edit. Our experiments comprised several loops including diverse dynamic variations. Results show that the methodology is scalable and facilitates diverse, novel, and cost-effective data collection. To our knowledge, the resulting dataset is the only expert-based multi-target HS/CN dataset available to the community.

The growing interest in employing counter narratives for hatred intervention brings with it a focus on dataset creation and automation strategies. In this scenario, learning to recognize counter narrative types from natural text is expected to be useful for applications such as hate speech countering, where operators from non-governmental organizations are supposed to answer to hate with several and diverse arguments that can be mined from online sources. This paper presents the first multilingual work on counter narrative type classification, evaluating SoTA pre-trained language models in monolingual, multilingual and cross-lingual settings. When considering a fine-grained annotation of counter narrative classes, we report strong baseline classification results for the majority of the counter narrative types, especially if we translate every language to English before cross-lingual prediction. This suggests that knowledge about counter narratives can be successfully transferred across languages.

Since state-of-the-art approaches to offensive language detection rely on supervised learning, it is crucial to quickly adapt them to the continuously evolving scenario of social media. While several approaches have been proposed to tackle the problem from an algorithmic perspective, so to reduce the need for annotated data, less attention has been paid to the quality of these data. Following a trend that has emerged recently, we focus on the level of agreement among annotators while selecting data to create offensive language datasets, a task involving a high level of subjectivity. Our study comprises the creation of three novel datasets of English tweets covering different topics and having five crowd-sourced judgments each. We also present an extensive set of experiments showing that selecting training and test data according to different levels of annotators’ agreement has a strong effect on classifiers performance and robustness. Our findings are further validated in cross-domain experiments and studied using a popular benchmark dataset. We show that such hard cases, where low agreement is present, are not necessarily due to poor-quality annotation and we advocate for a higher presence of ambiguous cases in future datasets, in order to train more robust systems and better account for the different points of view expressed online.

Emotion analysis in polarized contexts represents a challenge for Natural Language Processing modeling. As a step in the aforementioned direction, we present a methodology to extend the task of Aspect-based Sentiment Analysis (ABSA) toward the affect and emotion representation in polarized settings. In particular, we adopt the three-dimensional model of affect based on Valence, Arousal, and Dominance (VAD). We then present a Brexit scenario that proves how affect varies toward the same aspect when politically polarized stances are presented. Our approach captures aspect-based polarization from newspapers regarding the Brexit scenario of 1.2m entities at sentence-level. We demonstrate how basic constituents of emotions can be mapped to the VAD model, along with their interactions respecting the polarized context in ABSA settings using biased key-concepts (e.g., “stop Brexit” vs. “support Brexit”). Quite intriguingly, the framework achieves to produce coherent aspect evidences of Brexit’s stance from key-concepts, showing that VAD influence the support and opposition aspects.

Recently research has started focusing on avoiding undesired effects that come with content moderation, such as censorship and overblocking, when dealing with hatred online. The core idea is to directly intervene in the discussion with textual responses that are meant to counter the hate content and prevent it from further spreading. Accordingly, automation strategies, such as natural language generation, are beginning to be investigated. Still, they suffer from the lack of sufficient amount of quality data and tend to produce generic/repetitive responses. Being aware of the aforementioned limitations, we present a study on how to collect responses to hate effectively, employing large scale unsupervised language models such as GPT-2 for the generation of silver data, and the best annotation strategies/neural architectures that can be used for data filtering before expert validation/post-editing.

In the context of chit-chat dialogues it has been shown that endowing systems with a persona profile is important to produce more coherent and meaningful conversations. Still, the representation of such personas has thus far been limited to a fact-based representation (e.g. “I have two cats.”). We argue that these representations remain superficial w.r.t. the complexity of human personality. In this work, we propose to make a step forward and investigate stance-based persona, trying to grasp more profound characteristics, such as opinions, values, and beliefs to drive language generation. To this end, we introduce a novel dataset allowing to explore different stance-based persona representations and their impact on claim generation, showing that they are able to grasp abstract and profound aspects of the author persona.

Although there is an unprecedented effort to provide adequate responses in terms of laws and policies to hate content on social media platforms, dealing with hatred online is still a tough problem. Tackling hate speech in the standard way of content deletion or user suspension may be charged with censorship and overblocking. One alternate strategy, that has received little attention so far by the research community, is to actually oppose hate content with counter-narratives (i.e. informed textual responses). In this paper, we describe the creation of the first large-scale, multilingual, expert-based dataset of hate-speech/counter-narrative pairs. This dataset has been built with the effort of more than 100 operators from three different NGOs that applied their training and expertise to the task. Together with the collected data we also provide additional annotations about expert demographics, hate and response type, and data augmentation through translation and paraphrasing. Finally, we provide initial experiments to assess the quality of our data.

We present a novel abstraction framework called FASTDial for designing task oriented dialogue agents, built on top of the OpenDial toolkit. This framework is meant to facilitate prototyping and development of dialogue systems from scratch also by non tech savvy especially when limited training data is available. To this end, we use a generic and simple frame-slots data-structure with pre-defined dialogue policies that allows for fast design and implementation at the price of some flexibility reduction. Moreover, it allows for minimizing programming effort and domain expert training time, by hiding away many implementation details. We provide a system demonstration screencast video in the following link: https://vimeo.com/329840716

End-to-end neural approaches are becoming increasingly common in conversational scenarios due to their promising performances when provided with sufficient amount of data. In this paper, we present a novel methodology to address the interpretability of neural approaches in such scenarios by creating challenge datasets using dialogue self-play over multiple tasks/intents. Dialogue self-play allows generating large amount of synthetic data; by taking advantage of the complete control over the generation process, we show how neural approaches can be evaluated in terms of unseen dialogue patterns. We propose several out-of-pattern test cases each of which introduces a natural and unexpected user utterance phenomenon. As a proof of concept, we built a single and a multiple memory network, and show that these two architectures have diverse performances depending on the peculiar dialogue patterns.

We present a domain portable zero-shot learning approach for entity recognition in task-oriented conversational agents, which does not assume any annotated sentences at training time. Rather, we derive a neural model of the entity names based only on available gazetteers, and then apply the model to recognize new entities in the context of user utterances. In order to evaluate our working hypothesis we focus on nominal entities that are largely used in e-commerce to name products. Through a set of experiments in two languages (English and Italian) and three different domains (furniture, food, clothing), we show that the neural gazetteer-based approach outperforms several competitive baselines, with minimal requirements of linguistic features.

In task-oriented conversational agents, more attention has been usually devoted to assessing task effectiveness, rather than to how the task is achieved. However, conversational agents are moving towards more complex and human-like interaction capabilities (e.g. the ability to use a formal/informal register, to show an empathetic behavior), for which standard evaluation methodologies may not suffice. In this paper, we provide a novel methodology to assess - in a completely controlled way - the impact on the quality of experience of agent’s interaction strategies. The methodology is based on a within subject design, where two slightly different transcripts of the same interaction with a conversational agent are presented to the user. Through a series of pilot experiments we prove that this methodology allows fast and cheap experimentation/evaluation, focusing on aspects that are overlooked by current methods.

E-commerce platforms present products using titles that summarize product information. These titles cannot be created by hand, therefore an algorithmic solution is required. The task of automatically generating these titles given noisy user provided titles is one way to achieve the goal. The setting requires the generation process to be fast and the generated title to be both human-readable and concise. Furthermore, we need to understand if such generated titles are usable. As such, we propose approaches that (i) automatically generate product titles, (ii) predict their quality. Our approach scales to millions of products and both automatic and human evaluations performed on real-world data indicate our approaches are effective and applicable to existing e-commerce scenarios.

In this paper, we describe a methodology to infer Bullish or Bearish sentiment towards companies/brands. More specifically, our approach leverages affective lexica and word embeddings in combination with convolutional neural networks to infer the sentiment of financial news headlines towards a target company. Such architecture was used and evaluated in the context of the SemEval 2017 challenge (task 5, subtask 2), in which it obtained the best performance.

In computation linguistics a combination of syntagmatic and paradigmatic features is often exploited. While the first aspects are typically managed by information present in large n-gram databases, domain and ontological aspects are more properly modeled by lexical ontologies such as WordNet and semantic similarity spaces. This interconnection is even stricter when we are dealing with creative language phenomena, such as metaphors, prototypical properties, puns generation, hyperbolae and other rhetorical phenomena. This paper describes a way to focus on and accomplish some of these tasks by exploiting NgramQuery, a generalized query language on Google N-gram database. The expressiveness of this query language is boosted by plugging semantic similarity acquired both from corpora (e.g. LSA) and from WordNet, also integrating operators for phonetics and sentiment analysis. The paper reports a number of examples of usage in some creative language tasks.

The name of a company or a brand is the key element to a successful business. A good name is able to state the area of competition and communicate the promise given to customers by evoking semantic associations. Although various resources provide distinct tips for inventing creative names, little research was carried out to investigate the linguistic aspects behind the naming mechanism. Besides, there might be latent methods that copywriters unconsciously use. In this paper, we describe the annotation task that we have conducted on a dataset of creative names collected from various resources to create a gold standard for linguistic creativity in naming. Based on the annotations, we compile common and latent methods of naming and explore the correlations among linguistic devices, provoked effects and business domains. This resource represents a starting point for a corpus based approach to explore the art of naming.

Evaluating systems and theories about persuasion represents a bottleneck for both theoretical and applied fields: experiments are usually expensive and time consuming. Still, measuring the persuasive impact of a message is of paramount importance. In this paper we present a new ``cheap and fast'' methodology for measuring the persuasiveness of communication. This methodology allows conducting experiments with thousands of subjects for a few dollars in a few hours, by tweaking and using existing commercial tools for advertising on the web, such as Google AdWords. The central idea is to use AdWords features for defining message persuasiveness metrics. Along with a description of our approach we provide some pilot experiments, conducted both with text and image based ads, that confirm the effectiveness of our ideas. We also discuss the possible application of research on persuasive systems to Google AdWords in order to add more flexibility in the wearing out of persuasive messages.

In political speeches, the audience tends to react or resonate to signals of persuasive communication, including an expected theme, a name or an expression. Automatically predicting the impact of such discourses is a challenging task. In fact nowadays, with the huge amount of textual material that flows on the Web (news, discourses, blogs, etc.), it can be useful to have a measure for testing the persuasiveness of what we retrieve or possibly of what we want to publish on Web. In this paper we exploit a corpus of political discourses collected from various Web sources, tagged with audience reactions, such as applause, as indicators of persuasive expressions. In particular, we use this data set in a machine learning framework to explore the possibility of classifying the transcript of political discourses, according to their persuasive power, predicting the sentences that possibly trigger applause. We also explore differences between Democratic and Republican speeches, experiment the resulting classifiers in grading some of the discourses in the Obama-McCain presidential campaign available on the Web.

This paper presents resources and strategies for persuasive natural language processing. After the introduction of a specifically tagged corpus, some techniques for affective language processing and for persuasive lexicon extraction are provided together with prospective scenarios of application.

In this paper a first implementation of a tool for valence shifting of natural language texts, named Valentino (VALENced Text INOculator), is presented. Valentino can modify existing textual expressions towards more positively or negatively valenced versions. To this end we built specific resources gathering various valenced terms that are semantically or contextually connected, and implemented strategies that uses these resources for substituting input terms.