Dirk Hovy

🚨(Software) Update: In my PhD, I had a side project to fix an annoying problem: when you ask 5 people to label the same thing, you often get different answers. But in ML (and lots of other analyses), you still need a single aggregated answer. Using the majority vote is easy–but often wrong. 1/N

🌶️ @dirkhovy.bsky.social from our @milanlp.bsky.social lab will be joining today's #EMNLP2026 plenary panel on "New Missions in NLP Research"! Don't miss it! 🕝 14:30 📍 P1-3

Apart from an excuse to work with @questoph.bsky.social, this paper holds a special place in my heart for the funnest visualization (I love maps, and this paper has plenty):

The lab is well represented at EMNLP. I will also be in Budapest and am looking to recruit for an upcoming postdoc position. If you are there and interested (or if you just want to chat), ping me on Whova.

#MemoryMonday #NLProc 'Impoverished Language Technology: The Lack of (Social) Class in NLP' by Curry, Talat, @dirkhovy.bsky.social (2024). New tools target harmful content! #OnlineSafety arxiv.org/pdf/2403.03874

#MemoryModay #NLProc 'Two Contrasting Data Annotation Paradigms for Subjective NLP Tasks' by @paul-rottger.bsky.social et al. (2022) explores descriptive and prescriptive data labeling paradigms. aclanthology.org/2022.naacl-m...

#TBT #NLProc 'Welcome to the Modern World of Pronouns: Identity-Inclusive Natural Language Processing beyond Gender' by Anne Lauscher, Archie Crowley, Dirk Hovy (2022) discusses non-binary representation in tech and NLP's pronouns issue. aclanthology.org/2022.coling-...

#MemoryMonday #NLProc Hung et al.'s 2023 paper, 'Can Demographic Factors Improve Text Classification?' finds demographic adaptations of Transformer NLP models don't notably boost performance. aclanthology.org/2023.finding...

🌈 Join StereoQueerEval @ #SemEval2027 (Task 8)! Can #NLP systems detect stereotypes and hate speech targeting LGBTQIA+ people across English, Italian, Dutch & Persian? 📊 Training/dev data available now! 🗓️ Evaluation: Jan 10–31 👉 stereoqueereval.github.io/how_to_parti...

#TBT #NLProc 'Understanding Political Discourse on Twitter through Election Manifestos' Maurer et al. (2024) present a method for predicting party positioning from tweets aclanthology.org/2024.finding...

#TBT #NLProc  'Classist Tools: Social Class Correlates with Performance in NLP' by Curry et al. (2024) finds NLP tools perform worse on lower-SES speech - higher perplexity, more ASR errors, dialect flagged as "wrong" by grammar checkers. arxiv.org/pdf/2403.04445

Bocconi arranged for an interview with a student (Milek Anilkumar) on studying in the age of AI. It was a lot of fun, and could have easily been 4x longer. TLDR: still worth it… www.youtube.com/watch?v=kclp...

#MemoryModay #NLProc Countering Hateful and Offensive Speech Online - Open Challenges" (Flor Miriam Plaza-Del-Arco, @deboranozza.bsky.social, et al. 2024) presented a tutorial on the challenges and solutions for detecting and mitigating hate speech online. aclanthology.org/2024.emnlp-t...

📣 Call for Papers for the Student Research Workshop at NAACL 2027 Deadline: Jan. 11, 2027 This is a forum for student-led completed work or work in progress. It features archival and non-archival tracks. See the website for more info: naacl27-srw.github.io

🔔CFP for the #ACL2027NLP main conference is up! 2027.aclweb.org/calls/main/ 📝Submission deadline to ARR is Jan 4, 2027.

We are back with #TBT #NLProc Best Paper at NeurIPS D&B 2024: Hanna Rose Kirk, @paul-rottger.bsky.social et al. introduce 'The PRISM Alignment Dataset' arxiv.org/pdf/2404.16019

#MemoryModay #NLProc Outstanding Paper at ACL 2024! @paul-rottger.bsky.social et al. evaluate LLM values and opinions in 'Political Compass or Spinning Arrow?' aclanthology.org/2024.acl-lon...

#TBT #NLProc 'Understanding Political Discourse on Twitter through Election Manifestos' Maurer et al. (2024) present a method for predicting party positioning from tweets aclanthology.org/2024.finding...

#MemoryModay #NLProc #AI #MachineLearning #SafetyFirst 'Safety-Tuned LLaMAs: Improving LLMs Safety' by Bianchi et al. explores training LLMs for safe refusals, warns of over-tuning. arxiv.org/pdf/2309.07875

#TBT #NLProc 'Detecting Misogynous Memes with Text & Image Modalities' by Attanasio, @deboranozza.bsky.social, Bianchi. Their novel system uses Perceiver IO, surpassing all previous benchmarks. #AI #ScienceUpdate aclanthology.org/2022.semeval...

#MemoryModay #NLProc 'SocioProbe: What, When, and Where Language Models Learn about Sociodemographics' - Lauscher et al. delve into language models' grasp of sociodemographics. Their findings? Models learn, but don't apply it. aclanthology.org/2022.emnlp-m...

Our TACL paper "Let Guidelines Guide You: A Prescriptive Guideline-Centered Data Annotation Methodology" is finally out, only six months after acceptance! 😅🎉 direct.mit.edu/tacl/article...

#TBT #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.bsky.social et al. aclanthology.org/2022.emnlp-m...

Just arrived in lovely León. Tomorrow I’ll give a keynote at SEPLN, the Spanish Society for Natural Language Processing, on some of the most interesting problems ahead for NLP. But first: architecture and tapas! sepln2026.org/cuadro-de-po...

After Science News, our recent ICML paper also got featured in @science.org! Article: www.science.org/content/arti... Paper: arxiv.org/pdf/2605.03202 @dirkhovy.bsky.social @sanmikoyejo.bsky.social #science #PeerReview #research

Very happy that #IC2S22027 will be in Milan at Università Bocconi! I’m really looking forward to welcoming the computational social science community to my university and to Milan, and excited to be part of the team organizing it. See you in 2027! 🇮🇹✨

#MemoryModay #NLProc 'My Answer is C' by Wang et al. (2024) underscores the scrutiny needed for full text responses in LLMs multi-choice evaluations. aclanthology.org/2024.finding...

4134 submissions have been submitted to the ARR August cycle, a 14% increase compared to the ARR October 2025 cycle corresponding to EACL🔥. #EACL2027 in Athens is likely to be the biggest EACL ever 👀 #NLProc @aclrollingreview.bsky.social

#TBT #NLProc "Narratives at Conflict" by Sinelnik and @dirkhovy.bsky.social looks at hidden tactics of disinformation campaigns! They analyzed 8,000 news articles across 4 languages to reveal how disinformation campaigns adapt narratives for different audiences. 🕵️‍♀️ aclanthology.org/2024.acl-srw...

#TBT #NLProc 'Language is Scary when Over-Analyzed...' by @arimuti.bsky.social et al. explores argumentative reasoning in misogyny detection (2024). Detecting implicit misogyny proves challenging for language models. aclanthology.org/2024.emnlp-m...

#MemoryModay #NLProc Fornaciari et al.'s 2022 'Hard and Soft Evaluation of NLP models with BOOtSTrap SAmpling - BooStSa' is a useful Python tool for benchmarking NLP predictions with bootstrapped sampling. aclanthology.org/2022.acl-dem...

Text as Data has been a wonderful, long-running, non-archival workshop for empirical research at the intersection of AI and social science (especially work involving text). After a few years off, it will be happening again this year as a one-day event in early October!

🏆 Congrats to @zeerak.bsky.social and @dirkhovy.bsky.social on the 2016 Test of Time Award! 🎉 "Hateful Symbols or Hateful People? Predictive Features for Hate Speech Detection on Twitter" has shaped research on #hatespeech! Paper: aclanthology.org/N16-2013/ #NLProc

#TBT #NLProc 'Universal Joy: A Data Set and Results for Classifying Emotions Across Languages' by Lamprinidis et al. (2021) explores how AI research affects our planet. Tech can be green too! #SustainableTech aclanthology.org/2021.wassa-1.7

#TBT #NLProc 'Exploring challenges in Zero-shot Cross-lingual Hate Speech Detection, @deboranozza.bsky.social (2021) reveals how current models may inaccurately label non-hateful, language-specific interjections as hate speech signals.' aclanthology.org/2021.acl-sho...

#MemoryMonday #NLProc 'BERTective: Language Models and Contextual Information for Deception Detection' by Fornaciari, T. et al. (2021) explores AI's ability to detect deceit through context. www.aclweb.org/anthology/20...

#TBT #NLProc Hessenthaler et al.'s 2022 work delves into AI's link with fairness & energy reduction in English NLP models, challenging bias reduction theories. #AI #NLP #sustainability aclanthology.org/2022.emnlp-m...

Very happy to have given my first keynote at an #NLProc related conference! This morning at CORIA-TALN 2026 in Nantes, I talked about emerging risks and research directions around the everyday use of LLMs. coria-taln-2026.ls2n.fr/conferences-...

#MemoryMonday #NLProc 'Entropy-based Attention Regularization Frees Unintended Bias Mitigation from Lists' by Attanasio et al. redefines bias reduction in #AI, sans prior term knowledge. #2022Publication aclanthology.org/2022.finding...

#TBT #NLProc Bergman et al.'s 'Guiding the Release of Safer E2E Conversational AI through Value Sensitive Design' explores AI launch with a value-sensitive lens. aclanthology.org/2022.sigdial...

Thank you, @punarpuli.bsky.social, for a great article! Find our paper and more insights from Jiaxin Pei, Sanmi Koyejo, @dirkhovy.bsky.social, and me in this thread👇 bsky.app/profile/joac...

Could not be prouder of @zeerak.bsky.social for this accomplishment: from first paper ever (based on an MSc thesis) to 10-year ToT award. We could not have anticipated the lasting impact of this paper, but it's a great honor – and a wonderful story for any student/advisor team (see Zee's thread).

#MemoryMonday #NLProc Uma, A. N. et al. examine AI model training in 'Learning from Disagreement: A Survey'. Disagreement-handling methods' performance is shaped by evaluation methods & dataset traits. www.jair.org/index.php/ja...

@taniseceron.bsky.social is presenting her work about political content in pre-training and post-training data at the AI & Society conference. #AIandSociety #NLProc

On the heels of a fantastic Dagstuhl seminar on Social Intelligence in AI (thx, @jennhu.bsky.social, @maartensap.bsky.social, @tomerullman.bsky.social, & Lucie Flek), a callback to how we thought about this 5 years ago.

#TBT #NLProc 'Predictive Biases in Natural Language Processing Models: A Conceptual Framework and Overview' by Shah, Schwartz, Dirk Hovy (2020). Dive into their mathematical NLP bias framework, its origins, impacts, and types. #AIEthics aclanthology.org/2020.acl-mai...

#TBT #NLProc Check out 'Helpful or Hierarchical?' by Rashid, F. et al advocating better non-binary representation in tech. #Inclusion

aclanthology.org/2021.finding...

#MemoryModay #NLProc 'We will Reduce Taxes' - Identifying Election Pledges with Language Models' by Fornaciari et al. makes election promise tracking effortless with neural models.

#TBT #NLProc Bianchi, Terragni & Dirk Hovy present a smarter technique for topic modeling. Their method uses contextual embeddings for clear, meaningful word clusters, transforming how we interpret large text collections.