Report
Can vision models recognize objects without their usual neighbors?
An author says their NeurIPS 2026 paper examines typicality bias across modern vision models.
TLDR
An author of “Stranger Things: When Objects Appear Without Their Typical Neighbours” says the NeurIPS 2026 paper examines typicality bias across modern vision models. It asks whether a model can recognize an object when its usual neighbors disappear.
Combined views
2
1 Source, first seen ago
