Piotr studied computer science at ENSEEIHT and obtained his Ph.D. in deep learning at New York University under the supervision of Prof. Yann LeCun. He has a decade-long experience of machine learning in industrial research labs, where he developed solutions for epileptic seizure prediction from EEG, robotic navigation and natural language processing. His passion for performing arts, as a drama student with a 17-year background in improvised theatre, drew him to create HumanMachine, an artistic experiment fusing improv and AI, where Piotr’s alter-ego Albert shares the stage with a computer called A.L.Ex.
This talk and performance between Albert (a human) and A.L.Ex (an artificial intelligence embodied by a robot) from HumanMachine will showcase how improv comedy can be assisted (or made difficult) by a computer stage partner. The show was co-created by Piotr Mirowski and Kory Mathewson, two research scientists in artificial intelligence with a shared passion for building things as well as for drama and performing arts. The aim of the show is to build bridges between the artistic / acting and AI / hacker communities, and to raise questions about the nature of spontaneity and communication. Can we model language by patterns? How can a statistical language model inform improvised theatre, which is about doing the most "obvious" thing? How easily do we suspend disbelief? Why should we follow the basic improvisation principles of always saying "Yes, and..." and of making one's (stage) partner look good? The improv show will be in English, the following talk will be in French.