56 research outputs found
Human-in-the-Loop for Data Collection: a Multi-Target Counter Narrative Dataset to Fight Online Hate Speech
No full textUndermining 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
Human-Machine Collaboration Approaches to Build a Dialogue Dataset for Hate Speech Countering
No full textFighting 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.Comment: To appear in Proceedings of the 2022 Conference on Empirical Methods
in Natural Language Processing (long paper
Weigh your own words: improving hate speech counter narrative generation via attention regularization
No full textUsing Pre-Trained Language Models for Producing Counter Narratives Against Hate Speech: a Comparative Study
No full textIn 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
NLP for Counterspeech against Hate: A Survey and How-To Guide
No full textIn 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
Weigh Your Own Words: Improving Hate Speech Counter Narrative Generation via Attention Regularization
No full textRecent 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
Active feedback cooling of a SiN membrane resonator by electrostatic actuation
No full textFeedback-based control techniques are useful tools in precision measurements as they allow us to actively shape the mechanical response of high quality factor oscillators used in force detection measurements. In this paper, we implement a feedback technique on a high-stress low-loss SiN membrane resonator, exploiting the charges trapped on the dielectric membrane. A properly delayed feedback force (dissipative feedback) enables the narrowing of the thermomechanical displacement variance in a similar manner to the cooling of the normal mechanical mode down to an effective temperature Teff. In the experiment reported here, we started from room temperature and gradually increasing the feedback gain, we were able to cool down the first normal mode of the resonator to a minimum temperature of about 124mK. This limit is imposed by our experimental setup and, in particular, by the injection of the read-out noise into the feedback. We discuss the implementation details and possible improvements to the technique.Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.Electronic Components, Technology and MaterialsEKL Equipmen
Is Safer Better? The Impact of Guardrails on the Argumentative Strength of LLMs in Hate Speech Countering
No full textThe 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
Implementação de sistema de rastreabilidade em empresa de base tecnológica com ênfase em tecnologias da indústria 4.0
No full textWith the advent of Industry 4.0 and the approach of a connected production system using Industry 4.0 technologies, the adoption of traceability systems becomes an essential application of this new industrial era. The work takes into account the need for small and medium-sized companies to improve the processes carried out, through the use of technologies mentioned in the pillars of Industry 4.0. In this context, the work proposes to develop a traceability system applied to the production chain of stainless-steel capacitor tanks with the reading of the bar code in stainless steel AISI 304 and stainless steel AISI 409 of one of the units of a Brazilian company located in Itajubá, south of Minas Gerais. Initially, a systematic literature review conducted based on the PRISMA protocol was adopted, covering the Scopus and Web of Science databases, which allowed the selection of the most relevant articles on this topic. The methodology used was Action Research in which it was possible to develop and implement the proposed traceability system called RAST 4.0. Through a pilot batch, it was possible to collect data and analyze them through the stages of the PDCA cycle; thus, an Action Plan was developed with the purpose of solving the flaws found. Continuing with the action-research stages, with the application of the Action Plan, the company achieved improvements through simple actions, thus generating the 2nd Pilot Batch, which was monitored and tested within the process, resulting in time taken for each of tasks and maintaining traceability from the beginning to the shipment of the capacitor tank to the customer, through internal traceability.Com o advento da Indústria 4.0 e a abordagem de um sistema de produção conectada ao uso de tecnologias da Indústria 4.0, a adoção de sistemas de rastreabilidade vem ser uma aplicação essencial dessa nova era industrial. O trabalho leva em consideração a necessidade de empresas de pequeno e médio portes melhorarem os processos realizados, através do uso de tecnologias que constituem os pilares da Indústria 4.0. No contexto, o trabalho propõe desenvolver um sistema de rastreabilidade aplicado à cadeia de produção de tanque de capacitores de aço inox através da leitura do código de barras no aço inoxidável AISI 304 ou aço inoxidável AISI 409 de uma das unidades de uma empresa brasileira localizada em Itajubá, Sul de Minas Gerais. Inicialmente, adotou-se uma revisão sistemática de literatura conduzida com base no protocolo PRISMA, abrangendo as bases de dados Scopus e Web of Science, que permitiu a seleção dos artigos mais relevantes acerca dessa temática. O método de pesquisa utilizado foi a Pesquisa-ação na qual foi possível desenvolver e implantar o Sistema de rastreabilidade proposto chamado de RAST 4.0. Por meio de um lote piloto, foi possível coletar os dados e analisá-los através das etapas do ciclo PDCA; desse modo, um Plano de Ação foi desenvolvido com o propósito de solucionar as falhas encontradas. Dando prosseguimento nas etapas da pesquisa-ação, com a aplicação do Plano de Ação, a empresa conquistou melhorias através de ações simples. Foi gerado o 2º Lote Piloto, que foi acompanhado e testado dentro do processo, resultando em tomadas de tempo de cada uma das tarefas e mantendo a rastreabilidade desde o início até o envio do tanque de capacitor para o cliente, através da Rastreabilidade interna
Frequency-noise cancellation in optomechanical systems for ponderomotive squeezing
No full textPonderomotive squeezing of the output light of an optical cavity has been recently observed in the megahertz range in two different cavity optomechanical devices. Quadrature squeezing becomes particularly useful at lower spectral frequencies, for example, in gravitational wave interferometers, despite being more sensitive to excess phase and frequency noise. Here we show a phase and frequency-noise cancellation mechanism due to destructive interference which can facilitate the production of ponderomotive squeezing in the kilohertz range and we demonstrate it experimentally in an optomechanical system formed by a Fabry-P´erot cavity with a micromechanical mirror.MicroelectronicsElectrical Engineering, Mathematics and Computer Scienc
- …
