Tilburg center for Cognition and Communication (TiCC)

We study how people communicate with each other and how computer systems can be taught to communicate with us.

vignetweb

TiCC Colloquium: Emmanuel Dupoux

What: Towards developmental AI
Where: CZ 122
When: Wednesday, 4 October 2017, 12:45 - 13:45 hours


Abstract

Despite recent progress, Machine Learning is still not on par with the ability of humans of learning complex abilities such as language or common sense with little or no expert supervision. This talk presents recent work inspired by research in infant development on two topics: (1) The automatic discovery of linguistic units (phonemes, words) from raw speech using a variety of unsupervised or weakly supervised techniques, and (2) the automatic discovery of macroscopic physics from the observation of videos (the IntPhys Benchmark).

About Emmanuel Dupoux

E. Dupoux is full professor at the Ecole des Hautes Etudes en Sciences Sociales (EHESS), and directs the Cognitive Machine Learning team at the Ecole Normale Supérieure (ENS) in Paris and INRIA. His education includes a PhD in Cognitive Science (EHESS), a MA in Computer Science (Orsay University) and a BA in Applied Mathematics (Pierre & Marie Curie University, ENS). His research mixes developmental science, cognitive neuroscience, and machine learning, with a focus on the reverse engineering of infant language and cognitive development using unsupervised or weakly supervised learning. He is the recipient of an Advanced ERC grant, the founder of the Zero Resource Speech Challenge (2015, 2017) the Intuitive Physics Benchmark (2017) and led in 2017 a JHU Jelinek Memorial Summer Workshop at CMU on multimodal speech learning.

Website


When: 04 October 2017 12:45

End date: 04 October 2017 13:45