Workshop on reasoning & neural models

On May 19th we are organizing a small workshop on reasoning and neural network models. The focus will be on the question of whether neural networks can learn logical reasoning.  

The workshop is a satellite event of the CoSaQ Closing event.

Location: Room A1.24 at Science Park 904. The talks are open to the public.  

Schedule

13.30 – 14.15 Jelle Zuidema “How to teach neural networks logic, and how to demonstrate that they have learned it”
14.15 – 15:45 Maria Boritchev & Manuel Vargas Guzmán “Hybrid models of natural reasoning”
15.45 – 16.00 Coffee break
16.00 – 16.45 Jaap Jumelet “Inferring a Language Model’s Internal Grammar via Feature Interactions”
16.45 – 17.30 Dieuwke Hupkes “Are neural networks compositional, and how do we even measure that?”