Raphaël Millière

The Challenge of Compositionality for Artificial Intelligence

Official website

Recordings

Day 1: Why Compositionality Matters for AI

• Gary Marcus – "Compositionality and Natural Language Understanding"
• Allyson Ettinger – "Shades of Meaning Composition: Defining Compositionality Goals in NLU"
• Paul Smolensky – "Human-Level Intelligence Requires Continuous, Robustly Compositional Representations: Neurocompositional Computing for NECST-Generation AI"
• Raphaël Millière – "Compositionality Without Constituency"

Day 2: Can Language Models Handle Compositionality?

• Dieuwke Hupkes – "Are Neural Networks Compositional, and How Do We Even Know?"
• Tal Linzen – "Successes and Failures of Compositionality in Neural Networks for Language"
• Stephanie Chan – "Data Distributions Drive Emergent In-Context Learning in Transformers"
• Ellie Pavlick – "No One Metric is Enough! Combining Evaluation Techniques to Uncover Latent Structure"
• Brenden Lake – "Human-Like Compositional Generalization Through Meta-Learning"

Speakers

• Stephanie Chan (DeepMind)
• Allyson Ettinger (University of Chicago)
• Dieuwke Hupkes (European Laboratory for Learning and Intelligent Systems / Meta AI)
• Paul Smolensky (Johns Hopkins University/Microsoft Research Redmond)
• Brenden Lake (New York University / Meta AI)
• Tal Linzen (New York University / Google AI)
• Gary Marcus (New York University, Emeritus)
• Raphaël Millière (Columbia University)
• Ellie Pavlick (Brown University / Google AI)