RL Odyssey 4: Neuro-Symbolic Reinforcement Learning
Details
Reinforcement Learning (RL) is now one of the most successful solutions to the problem of autonomous agency in complex and heterogeneous environments. However, several gaps remain addressed towards its deployment in the real world. For instance, state-of-the-art algorithms are data-hungry and computational intensive, and black-box RL policies are often untrusted due to their stochastic nature and the dimension of neural models.
This talk will focus on a trending topic of research addressing the aforementioned issues, namely neurosymbolic RL.
Neurosymbolic RL combines standard data-driven RL with symbolic (e.g., logical) approaches, which have a long successful history in artificial intelligence and autonomous agency in particular. Different paradigms for neurosymbolic integration will be presented, with example use cases showing the enhanced performance of neurosymbolic RL, including generalizability and transparency.
L'evento inizia alle 18:00, il talk inizia alle 18:30.
Speaker: Daniele Meli
Aula: T04
REGISTRAZIONE:
https://www.eventbrite.it/e/888311663097