Preprint
Aug 2026
Secrets Everywhere: Auditing Memorization in Mobility Prediction Models
A framework to quantify mobility memorization at different levels of granularity: individual locations, anchor pairs, and subtrajectory segments is introduced and user-grounded reference sets are developed to assess how likely a model is to prefer training data over realistic alternatives.
A. J. Kouam, Hristo Boyadzhiev, Konrad Rieck
· 0 citations