Category: Meetings

Beatrice Savoldi will present one of her recent work
entitle “Hi Guys or Hi folks? Navigating Gender Bias and Inclusive Language in Translation Technologies”.

Abstract

Societal gender asymmetries and inequalities can be embedded in our communication practices and perpetuated in language technologies, including Machine Translation (MT) systems used as scale. In this presentation, we will delve into the current landscape of MT and gender bias, as well as current proposals towards more inclusive language. 

By focusing on English-Italian as an exemplar language pair, we will discuss the challenges and opportunities — both theoretical, technical but also linguistic —  in fostering a more equitable automatic translation. 

When: 14/03/2025 11.00

Where: Aula 3.06 Thin Client (terzo piano) – Via Sant’Ottavio 54

Horacio Jarquin will present one of his projects
entitle “GAttention: Gated Attention for the Detection of Abusive Language”.

Abstract

Abusive language online creates toxic environments and exacerbates social tensions, underscoring the need for robust NLP models to interpret nuanced linguistic cues. This research introduces GAttention, a novel Gated Attention mechanism that combines the strengths of Contextual attention and Self-attention mechanisms to address the limitations of existing attention models within the text classification task. GAttention capitalizes on local and global query vectors by integrating the internal relationships within a sequence (Self-attention) and the global relationships among distinct sequences (Contextual attention). This combination allows for a more nuanced understanding and processing of sequence elements, which is particularly beneficial in context-sensitive text classification tasks such as the case of abusive language detection. By applying this mechanism to transformer-based encoder models, we showcase how it enhances the model’s ability to discern subtle nuances and contextual clues essential for identifying abusive language, a challenging and increasingly relevant task within NLP.

When: 21/03/2025 11.30

Where: Sala Riunioni (first floor)

Adriana Pagano will present one of her international projects
entitle “Bridging views on the concepts of ‘multilingual’, ‘cross-lingual’ and ‘translingual’ in Language Technology”.

Abstract

The presentation will introduce the interdisciplinary research network UniDive (Universality, Diversity, and Idiosyncrasy in Language Technology), a COST Action (European Cooperation in Science and Technology). Adriana will focus on one of the tasks she is currently leading within UniDive’s Working Group 3 – Multilingual and Cross-Lingual Language Technology.

When: 17/01/2025 11.30

Where: Sala Conferenze (3rd floor)

Marco Madeddu is a PhD student at the beginning of her 1st year, and he will talk about one of his latest works called “Exploring YouTube Comments Reacting to Femicide News in Italian”

Abstract

In recent years, the Gender Based Violence (GBV) has become an important issue in modern society and a central topic in different research areas due to its alarming spread.
Several Natural Language Processing (NLP) studies, concerning Hate Speech directed against women, have focused on misogynistic behaviours, slurs or incel communities.
The main contribution of our work is the creation of the first dataset on social media comments to GBV, in particular to a femicide event.
Our dataset, named GBV-Maltesi, contains 2,934 YouTube comments annotated following a new schema that we developed in order to study GBV and misogyny with an intersectional approach.
During the experimental phase, we trained models on different corpora for binary misogyny detection and found that datasets that mostly include explicit expressions of misogyny are an easier challenge, compared to more implicit forms of misogyny contained in GBV-Maltesi.

When: 13/12/2024 11.30

Where: Sala Conferenze (3rd floor)

Soda Marem Lo is a PhD student at the beginning of her 3rd year, and she will talk about one of her latest works called “I’m sure you’re a real scholar yourself: Exploring Ironic Content Generation by Large Language Models.”

Join us if you’re interested in learning more about the ability of LLMs to generate ironic content!

Abstract:

Generating ironic content is challenging: it requires a nuanced understanding of context and implicit references and balancing seriousness and playfulness. Moreover, irony is highly subjective and can depend on various factors, such as social, cultural, or generational aspects. This paper explores whether Large Language Models (LLMs) can learn to generate ironic responses to social media posts. To do so, we fine-tune two models to generate ironic and non-ironic content and deeply analyze their outputs’ linguistic characteristics, their connection to the original post, and their similarity to the human-written replies. We also conduct a large-scale human evaluation of the outputs. Additionally, we investigate whether LLMs can learn a form of irony tied to a generational perspective, with mixed results

When: 8/11/2024 11.30

Where: Sala Riunioni (1st floor)

Tom Bourgeade will present his research on “Data Augmentation through Back-Translation for Stereotypes and Irony Detection”.

Abstract

In NLP, the detection of nuanced phenomena such as stereotypes or irony presents unique challenges, namely linked to the scarcity of labeled datasets. One strategy to mitigate this is to employ Data Augmentation methods, which each have their pros and cons with regard to these phenomena. This presentation will focus on Back-Translation, which proposes exploiting modern Machine Translation models to introduce variety in training instances, in a process similar to paraphrasing, by translating a text into a pivot language, then back into the original language. We compare this approach on multilingual datasets for stereotypes and irony detection, against simpler strategies such as oversampling, as well as Cross-Translation, in which instances from other language subsets are translated and injected into the target training language subset.

When: 19/04/2024 11.30

Where: Sala Conferenze (3rd floor)

[trigger warning: mentions of gender-based violence]

When talking about societally impactful events, our choices of words and grammatical constructions often reflect our socio-political perspective on these events and affect how the people that we talk to perceive the events. In particular, in events that involve an unequal power relationship between different groups of people, this relationship affects how the agency of the participants in the events is portrayed. Gender-based violence is a particularly relevant example of this: “woman tragically dies in family incident” and “man suspected of killing his wife” could both be factually accurate descriptions of a femicide, but, when used as a newspaper headline, convey very different views of the event. In the lecture, we will discuss ways in which recently developed NLP techniques can help make visible such different ‘framings’ and contribute to increasing societal awareness. The lecture will be followed by a hands-on session in which we will do small-scale experiments together, looking at how to apply and extend these techniques.

Bio:
Gosse Minnema is a computational linguist based in Groningen, The Netherlands. He is currently preparing to defend his PhD thesis on frame semantics applied to media framing. His main area of interest is computational semantics and, in a broad sense, ways of applying it in societally meaningful ways. He is currently also an active member of the project “PeARS: The People’s Search Engine” (https://pearsproject.org/) which aims to promote community-owned, privacy-friendly and sustainable NLP solutions for web search and knowledge management.

When: March 26, h 14
Where: Via Sant’Ottavio 54, Room 3.06

Beatrice Savoldi will come to Turin for a seminar / lesson as part of the course of Ethics in NLP (Master degree in Language Technologies and Digital Humanities).
Feel free to join if you are interested in the topic!

Where: Aula 3.06 Thin Client (terzo piano) – Via Sant’Ottavio 54 – Via Sant’Ottavio – 54 – Torino

Title: Hi Guys or Hi folks? Navigating Gender Bias and Inclusive Language in Translation Technologies.

When: 11:00-12:30 March 15, 2024

Abstract:  Societal gender asymmetries and inequalities can be embedded in our communication practices and perpetuated in language technologies, including Machine Translation (MT) systems used as scale. In this presentation, we will delve into the current landscape of MT and gender bias, as well as current proposals towards more inclusive language.
By focusing on English-Italian as an exemplar language pair, we will discuss the challenges and opportunities — both theoretical, technical but also linguistic —  in fostering a more equitable automatic translation.

Short bio: Beatrice Savoldi is a postdoc researcher at Fondazione Bruno Kessler (FBK) within the MT research unit, where she mainly works
on trustworthy and gender inclusive translation technologies. Beatrice carried out a joint international PhD at the University of Trento
and Augsburg with a dissertation on gender bias in Machine Translation, which was awarded the 2023 best thesis Research Prize
from the Augsburg University Foundation. Her research interests broadly encompass ethical and social considerations of language technologies.

Massimo Donini will present his work “Multimodal Strategies for Robot-to-Human Communication”.

Abstract

Multimodality offers new possibilities in the field of robot-to-human communication. In the proposed approach, the coordinated  and integrated use of multimedia elements with the robot’s speech  plays a very important role in the overall effectiveness of the communicative act. During the research, different multimodal communication strategies have been formalised and implemented.

When: 09/02/2024 11:30

Where: Sala Riunioni (1st floor)

Where: Sala conferenze (3th floor)
When: January 26, 2024, 11:30

Luca Molinaro is a PhD student from our 39th cycle, and he will present his work titled ‘DelBERTo: A Deep Lightweight Transformer for Sentiment Analysis,’ which has been accepted at AIxIA 2022. Please find the abstract below.

Abstract:  
This article introduces DelBERTo, a resource-efficient Transformer architecture for Natural Language Processing (NLP). Transformers replace convolutions and recurrence with the self-attention mechanism and represent the state-of-the-art in NLP. However, self-attention’s complexity grows quadratically with the size of the input, which limits their applications. DelBERTo relies on adaptive input and on a deep yet lightweight Transformer architecture to reduce the number of learnable parameters, and relies on adaptive softmax to improve pre-training speed and memory footprint. We evaluate the proposed architecture in a sentiment analysis task and compare it against AlBERTo, a BERT model representing the state-of-the-art in sentiment analysis over Italian tweets. DelBERTo has only one-seventh of AlBERTo’s learnable parameters, is faster, and requires less memory. Despite this, our experiments show that DelBERTo is competitive with AlBERTo over the three SENTIPOLC sub-tasks proposed at EVALITA 2016: subjectivity classification, polarity classification, and irony detection.