TopoBrick: Agentic Topology Sampling of Exogenous Variables for Zero-Shot Building IoT Forecasting

2026-07-07

AI Summary

Existing building IoT forecasting methods often neglect the physical topology and contextual dependencies among sensors, either processing time series in isolation or relying on fixed covariates. To address this limitation, this work proposes TopoBrick—a zero-shot, training-free prediction framework that introduces, for the first time, an agent-based, topology-aware exogenous variable sampling mechanism. TopoBrick leverages a building knowledge graph to construct a structural skeleton and dynamically selects both historical and future-known exogenous variables tailored to each target sensor. Evaluated on three real-world building datasets, TopoBrick substantially outperforms strong zero-shot baselines and achieves performance comparable to specialized trained models. Ablation studies confirm the efficacy of the topology-aware sampling strategy, demonstrating clear advantages over random or fixed selection approaches.

Abstract

Building sensors are embedded in physical topology, spatial hierarchy, and operational context, yet existing forecasters often treat them as isolated time series or rely on fixed covariate sets. We present TopoBrick, a training-free framework for zero-shot building IoT (Internet-of-Things) forecasting. TopoBrick uses building knowledge graphs to construct a compact structural skeleton and employs an agentic topology sampler to select target-specific exogenous variables. The selected variables are organized by deployment-time availability, separating past-known sensor states from future-known calendar, schedule, and meteorological exogenous variables. Across three real-world buildings, TopoBrick outperforms strong zero-shot foundation-model baselines and remains competitive with fully trained building-specific models. Ablations show that topology-aware sampling is more reliable than random, ontology-only, or fixed-hop selection, especially for physically coupled HVAC and weather-driven sensing variables.

Problem

• building IoT forecasting
• zero-shot learning
• exogenous variables
• physical topology
• sensor time series

Innovation

• zero-shot forecasting
• topology-aware sampling
• building knowledge graph
• exogenous variable selection
• agent-based sampling