Safety Architecture
Safety architecture research focuses on designing robust systems to mitigate risks associated with increasingly autonomous technologies, particularly in AI agents, automated vehicles, and human-robot interaction. Current efforts concentrate on developing layered architectures incorporating input filtering, safety agents, and hierarchical delegation, alongside formal verification methods to ensure safety properties are met. This work is crucial for enabling the safe deployment of advanced technologies across various sectors, improving both system reliability and user trust.
Papers
December 12, 2024
September 3, 2024
August 8, 2024
October 20, 2023
September 18, 2023