Join Zoom Meeting
https://us02web.zoom.us/j/88276361301?pwd=S1c2UnlWRG9sRk1icVhxU2xmZnpIZz09

Meeting ID: 882 7636 1301
Passcode: 446189

Recent deep learning models are starting to demonstrate human-level abilities on an increasingly broader range of tasks spawning domains such as language understanding, reasoning, vision and decision making. The talk will provide an introduction to the core technologies and technical trends driving these breakthroughs. It will then turn to some of the more unexpected and interesting properties these models exhibit, along with their current shortcomings. Finally we touch upon what future developments of deep learning models could look like. The aim of the talk is to provide the background understanding necessary to assessing the societal and ethical implications of the large-scale deployment of artificial intelligence.

*

Florin Gogianu is a researcher at Bitdefender, with an initial background in philosophy, currently doing a PhD in deep reinforcement learning at the Technical University of Cluj-Napoca, under the supervision of prof. Lucian Bușoniu.