Search papers, labs, and topics across Lattice.
This study investigates the privacy risks associated with structured representations exported from domestic robots, revealing that even local sensor data can leak sensitive household information. The authors introduce Task-Functional Perception Distillation (TFPD), a framework that evaluates the privacy implications of perception exports based on task utility and exposure risks. Their experiments demonstrate that while navigation tasks can achieve high success rates, the linkability of representations varies significantly, highlighting the need for task-specific privacy assessments in robotic perception systems.
Task-specific evaluations reveal that conventional abstraction methods fail to provide consistent privacy guarantees in robot perception exports.
Domestic robots rely on rich perception to operate in private homes, but privacy risk persists even when raw sensor data remain local. Structured representations exported to downstream planners, cloud services, logs, or learning pipelines can still reveal household information through semantics, geometry, spatial structure, and task targets. We introduce Task-Functional Perception Distillation (TFPD), a task-scoped representation-export framework that keeps rich perception local and profiles downstream exports according to task utility, direct exposure, and multiple residual inference risks. Using 120 AI2-THOR scenes with scene-disjoint train/validation/test splits, frozen attacker selection, and representation-aware held-out attacks, we evaluate navigation, collision checking, and object-goal execution. Three navigation exports achieve identical success (1.000) and mean path ratio (0.898), yet representation-level linkability ranges from 0.532 to 0.970. Replacing an explicit target label with a target region reduces target-category macro-F1 from 1.000 to 0.077 while preserving success at 0.995, while geometric coarsening reduces object-category macro-F1 from 0.704 to 0.556 at a measurable collision-utility cost. A ProcTHOR replication preserves the navigation task-equivalence/privacy-inequivalence finding while changing the relative ordering of normalized and topological exports. These results show that neither field removal nor stronger abstraction induces a universal privacy ordering and motivate task-specific, multi-risk evaluation of the complete public representation.
