How undesired goals can arise with correct rewardsRohin Shah, Victoria Krakovna, Vikrant Varma, Zachary KentonOctober 7, 2022 “As we build increasingly advanced artificial intelligence (AI) systems, we want to make
What I Read: Exploring Virtual Worlds, AI
https://www.quantamagazine.org/ai-makes-strides-in-virtual-worlds-more-like-our-own-20220624/ By Exploring Virtual Worlds, AI Learns in New WaysAllison WhittenContributing WriterJune 24, 2022 “Intelligent beings learn by interacting with the world. Artificial intelligence researchers have adopted a similar strategy
What I Read: Policy Regulariser, Adversary
https://deepmindsafetyresearch.medium.com/your-policy-regulariser-is-secretly-an-adversary-14684c743d45 Your Policy Regulariser is Secretly an AdversaryDeepMind Safety ResearchMar 24 By Rob Brekelmans, Tim Genewein, Jordi Grau-Moya, Grégoire Delétang, Markus Kunesch, Shane Legg, Pedro A. Ortega“Policy regularisation can be
What I Read: Neural-Control Family
https://www.gshi.me/blog/NeuralControl/ Neural-Control Family: What Deep Learning + Control Enables in the Real WorldGuanya Shi “…is machine learning (especially deep learning) really ready to be deployed in safety-critical systems?”