#MemoryMonday #NLProc 'Exploring challenges in Zero-shot Cross-lingual Hate Speech Detection, @debora_nozza (2021) reveals how current models may inaccurately label non-hateful, language-specific interjections as hate speech signals.'

Last week we held our 1st MilaNLP retreat by beautiful Lago Maggiore! ��️🌊.We shared research ideas, stories (academic & beyond), and amazing food. It was a great time to connect outside of the usual lab working days, and most importantly, strengthen our bonds as a team. #NLProc

#TBT #NLProc 'Emotion Analysis in NLP: Trends, Gaps & Future' by Plaza-del-Arco et al. (2024) analyses key NLP papers, spotting major gaps in emotion analysis including terminology standardization and demographic aspects.

#MemoryMonday #NLProc 'Language Invariant Properties in NLP' by Bianchi, @debora_nozza, Dirk Hovy (2022) explores text transformation constants, studying translation & paraphrasing effects on sentiment & traits. #linguistics

🎓For today's lab seminar, we've had the pleasure to listen to @pietro_lesci 's presentation on "The hidden cost of split: Tokenization bias in language models". #NLProc

#TBT #NLProc Metrics for What, Metrics for Whom: Assessing Actionability of Bias Evaluation Metrics in NLP" by Delobelle, Attanasio, @deboranozza.bsky.social, Blodgett, and Talat, 2024 highlights the need for clearer, actionable bias measures in NLP.

📖For this Thursday's Reading Group, Yujie Ren presented ."How Does Vision-Language Adaptation Impact the Safety of Vision Language Models?" by Lee et al 2024. Paper: .#NLProc #AISafety

#MemoryMonday #NLProc 'Data-Efficient Strategies for Expanding Hate Speech Detection into Under-Resourced Languages' funnels limited target-language data to fine-tune models & enhance effectiveness. By @paul_rottger et al.

🎓For today's lab seminar it was a pleasure to have Federico Ruggeri presenting "From data annotation to modelling, the importance of integrating and extracting textual knowledge in hate speech" . #NLProc

#TBT #NLProc New study, 'Entropy-based Attention Regularization Frees Unintended Bias Mitigation from Lists' by Attanasio et al. redefines bias reduction in #AI, sans prior term knowledge. #2022Publication

📖For this Thursday's Reading Group @taniseceron.presented "Generative Echo Chamber? Effects of LLM-Powered Search Systems on Diverse Information Seeking " by Nikhil Sharma, Q. Vera Liao and Ziang Xiao. Paper: . #NLProc

#MemoryMonday #NLProc 'Interpretability Informs Gender Bias Mitigation in Machine Translation'- Attanasio et al. (2023). Their new research tackles gender bias in machine translations using a simple, few-shot learning method for fairness.

🎓For last Friday's lab seminar, it was a pleasure to have @shehzaadzd presenting "Learning to Adapt and Align with Feedback" . #NLProc

#TBT #NLProc 'Impoverished Language Technology: The Lack of (Social) Class in NLP' by Curry, Talat, Dirk Hovy (2024). New tools target harmful content! #OnlineSafety

📖For this Thursday's Reading Group @anne_lauscher .presented "Aligned but Blind: Alignment Increases Implicit Bias by Reducing Awareness of Race " by Lihao Sun, Chengzhi Mao, Valentin Hofmann and Xuechunzi Bai. Paper: #NLProc

#MemoryMonday #NLProc 'Visualizing Regional Language Variation Across Europe on Twitter' by Dirk Hovy et al. uncovers language differences across Europe with stunning visuals. Language is art! #Linguistics

🎓For today's lab seminar, it was a pleasure to have.@ZhijingJin presenting "Testing Moral Reasoning in Large Language Models: Trolley Problem, Political Bias, and Beyond" . #NLProc

RT @BeatriceSavoldi: 🔍 Ciao! Stiamo studiando come l'AI viene usata in Italia e per farlo abbiamo costruito un sondaggio!. 👉.

#TBT #NLProc Soni et al.'s (2024) "Comparing Pre-trained Human Language Models" shows language isn't just words-it's about the PERSON writing them 👤 However, current LLMs miss crucial human context. One-size-fits-all approaches don't work!

#MemoryMonday #NLProc 'Importance of Modeling Social Factors of Language: Theory and Practice' by Dirk Hovy, Diyi Yang discusses NLP limitations, calling for focus on social factors, not just content.