Stealthy World Model Manipulation Via Data Poisoning
Researchers have developed a method for stealthy world model manipulation via data poisoning. This technique targets model-based learning agents that use world models to predict future states and plan actions.
Topics
Developing
- 898d Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore.
- 898d Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur.
- 898d Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est.
- 898d Sed ut perspiciatis unde omnis iste natus error sit voluptatem accusantium doloremque laudantium.
Sources · 7 independent
Modernity/arxiv
“Stealthy World Model Manipulation via Data Poisoning. Authors: Yibin Hu, Xiaolin Sun, Zizhan Zheng Abstract: Model-based learning agents use learned world models to predict future states, plan actions, and adapt to new environments.”
Unlock the full story
Get a Pro subscription or above to see the live story progression and the full list of independent sources confirming each event as they happen.
Log in to upgrade