RL Odyssey 5: What's next for embodied AI?

n recent years, we have witnessed a remarkable improvement in the broader field of AI due to large-scale datasets and high-capacity (foundational) models. This progress has enabled a paradigm shift in the Embodied AI field, where agents learn through interaction and exploration to solve challenging tasks within their environment. This talk will introduce the common tasks of embodied AI, such as object finding (e.g., "find the black sofa with two pillows on it"), visual navigation (e.g., "go to the kitchen, then turn left at the first door and stop there"), and map representation, as well as the recent approaches using the latest advancements in foundational models.

Event Venue & Nearby Stays

Ca' Vignal 3, Strada le Grazie, Verona, VR, Italy