AI Misgeneralization Risks
Alan and Tim discuss the risks of misgeneralization in artificial intelligence, particularly in reinforcement learning systems. They delve into the potential failures in training goals translating to test time outcomes, highlighting the need for evaluation and solutions as models scale up.In this clip
From this podcast
Machine Learning Street Talk (MLST)
#94 - ALAN CHAN - AI Alignment and Governance #NEURIPS
Related Questions