Transformers for Decision-Making: From Language Models to Trajectory Optimization.
Transformers for Decision-Making: From Language Models to Trajectory Optimization.
Speaker: Daniele Gammelli (Standford's Autonomous Systems Lab (ASL))
Chair: Davide Cacciarelli
Date: 24th April 2024, at 17:00-18:00 CEST
This webinar will introduce core concepts regarding modern high-capacity neural network models (i.e., Transformers), trajectory optimization, and how these can be combined for effective decision-making in safety-critical tasks. By the end of the talk, we will aim to provide the audience with the conceptual tools needed to get started in the field of foundation models for control, and how to apply them in practice to achieve reliable robot autonomy.
Bio:
Daniele Gammelli is a postdoctoral scholar in Stanford’s Autonomous Systems Lab, where he focuses on developing learning-based solutions that enable the deployment of future autonomous systems in complex environments, with an emphasis on large-scale robotic networks, mobility systems and autonomous spacecraft. He received his Ph.D. in Machine Learning and Mathematical Optimization at the Technical University of Denmark, where he developed ML-based solutions to analyze and control future Intelligent Transportation Systems.
