Researchers have introduced TopoBrick, a novel framework designed for zero-shot forecasting in building Internet of Things (IoT) environments. This training-free system leverages building knowledge graphs to create a structural skeleton and an agentic topology sampler to identify relevant exogenous variables. TopoBrick organizes these variables based on their availability at deployment time, distinguishing between past sensor states and future-known data like weather and schedules. In tests across three real-world buildings, TopoBrick demonstrated superior performance compared to existing zero-shot foundation models and rivaled fully trained, building-specific models. AI
IMPACT This framework could improve the accuracy and efficiency of forecasting for smart buildings by better utilizing sensor data and external factors.
RANK_REASON The cluster contains a research paper detailing a new framework for IoT forecasting.
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
