Skip to content
Sign in
Programme

Coffee and posters

Sun 18 Oct · 11:30–12:30

Posters

Main foyer. Posters stay up all day.

Papers

Sign in to ask or upvote questions. Sign in

  1. Poster

    Privacy-preserving occupancy detection from smart-meter data at the building edge

    Nadia Rahman (Thalassa University)

    Abstract

    Smart meters reveal when people are home. We detect room occupancy for heating control on a gateway inside the building, so raw readings never leave it, and share only aggregated, differentially private statistics with the utility. Detection F1 is 0.88 and the utility's load forecast loses 2% accuracy compared with full data access.

    smart meters · occupancy · differential privacy · edge AI

    Questions from the audience 0

    No questions yet.