Forschungsinstitutionen

Join Zoom Meeting
https://zoom.us/j/99715655535?pwd=ZnVlTXdrcUgwREo2S1NVcG5wcmY4QT09

Meeting ID: 997 1565 5535
Passcode: jls-21

---------------------------------------------------------------

Abstract:

Indisputably, the Web has revolutionized the way
people receive, consume, and interact with information.
At the same time, unfortunately, the Web offers a
fertile ground for amplifying socio-technical issues
like the spread of hateful content and false information;
hence there is a pressing need to develop techniques
and tools to understand, detect, and mitigate these
issues on the Web. In this talk, I will present our
data-driven quantitative approach towards understanding
such emerging socio-technical problems, focusing on
the spread of hateful content targeting specific
demographic groups. Also, I will present our work on
understanding mitigation strategies employed by social
networks (i.e., moderation interventions) and assessing
their effectiveness towards reducing the prevalence
and impact of these issues on the Web.

Presenter:

Savvas Zannettou: https://www.mpi-inf.mpg.de/de/departments/inet/people/savvas-zannettou

--------------------------------------------------------------

Please find all speakers for this semester here:
https://nextcloud.mpi-klsb.mpg.de/index.php/s/LTNKiGWZ8e688kp

Join Zoom Meeting

https://zoom.us/j/99457028566?pwd=bGllck5vS05iYnhtMGVCUzFGS1JqUT09

Meeting ID: 994 5702 8566
Passcode: 780865

Abstract:

Distributional reinforcement learning (RL) is a new approach to RL with
the emphasis on the distribution of the rewards obtained rather than just
the expected reward as in traditional RL. In this work we take the
distributional point of view and analyse a number of sampling-based
algorithms such as value iteration, TD(0) and policy iteration. These
algorithms have been shown to converge under various assumptions but
usually with completely different proofs. We have developed a new
viewpoint which allows us to prove convergence using a uniform approach.
The idea is based on couplings and on viewing the approximation algorithms
as Markov processes in their own right. It originated from work on
bisimulation metrics in which I have been working for the last quarter
century. This is joint work with Philip Amortila (U. Illinois), Marc
Bellemare (Google Brain) and Doina Precup (McGill, Mila and DeepMind).

Speaker's bio:

Prakash Panangaden is a Professor of Computer Science at McGill University.
His research interests are primarily in theoretical foundations of computer science
with a focus on stochastic systems, but ranges from black holes and curved space-time
to reinforcement learning. He has received numerous awards, including the
Test of Time award at LICS. He is a Fellow of the ACM.

Please contact the host directly (rupak@mpi-sws.org) if you would like to meet with Prakash Panangaden.