Coding Aperitivo

The Coding Aperitivo is our take on a weekly seminar series. We end the working week and wind down with some relaxed academic chatter, a drink and some snacks.


We usually host external speakers on Fridays at 4pm Milan time. Talks are mostly virtual, and sometimes in person. We encourage our guests to try different formats with us such as guided discussions, hands-on activities, debates or just a nice academic chat. Got some research ideas you and you want a sounding board? We are very happy to discuss ongoing or upcoming research.

Past Guests


  • Emily Sheng: “Biases in NLG and Dialogue Systems”
  • Nedjma Ousidhoum: “Expectations vs. Reality when Working on Toxic Content Detection in NLP”
  • Nils Reimers: “Training State-of-the-Art Text Embedding & Neural Search Models”
  • Maarten Sap: “Detecting and Rewriting Socially Biased Language”
  • Sunipa Dev: “Towards Interpretable, Fair and Socially-Aware of Language Representations”
  • Alba Curry: “Philosophy of Emotion and Sentiment Detection”
  • Rob van der Goot: “Multi-lingual and Multi-task learning: from Dataset Creation to Modeling
  • Su Lin Blodgett: “Social and Ethical Implications of NLP Technologies”
  • Gabriele Sarti: “Interpreting Neural Language Models for Linguistic Complexity Assessment”
  • Paul Röttger :Two Contrasting Data Annotation Paradigms for Subjective NLP Tasks”
  • Chia-Chien Hung: “Multi-domain and Multilingual Dialog”
  • Anna Wegmann: “Does It Capture STEL? A Modular, Similarity-based Linguistic Style Evaluation Framework”
  • Abhilasha Ravichander: “Probing the Probing Paradigm: Does Probing Accuracy Entail Task Relevance?”
  • Samson Tan (AWS AI Research & Education): “Towards Sociolinguistically-Inclusive NLP: An Adversarial Approach”




  • Emanuele La Malfa: “Code Simulation Challenges for Large Language Models”
  • Enrico Liscio: “Context-Specific Value Inference via Hybrid Intelligence”
  • Eve Fleisig: “When the Majority is Wrong: Modeling Annotator Disagreement for Language Tasks”
  • Vishakh Padmakumar: “Does Writing with Language Models Reduce Content Diversity?”
  • Enrico Bertino: “AI at a Milanese Chatbot Start-Up”
  • Fangru Lin: “Graph-enhanced Large Language Models in Asynchronous Plan Reasoning”

Dirk’s Negroni

When in Milan, drink as the Milanese. Though the official recipe calls for equal parts gin, Campari and red vermouth, here we opt for a punchier negroni, heavy on the gin. For a sbagliato, substitute the gin for prosecco.

  • 3-4 parts gin (to taste)
  • 2 parts campari
  • 1 part red vermouth.

Pour the ingredients into a mixing glass with ice and stir until the glass feels very cold. Strain into a glass with a (very) large ice cube and a twist of orange (and rub the glass rim with it).