InsightBuild: LLM-Powered Causal Reasoning in Smart Building Systems

2025-07-10

AI Summary

Frequent anomalous energy consumption in smart buildings often remains unexplained to operations personnel. To address this, we propose an interpretable causal analysis framework that—uniquely—integrates explicit causal discovery with large language models (LLMs) to enable end-to-end generation of natural-language causal explanations from heterogeneous sensor time-series data. Methodologically, we employ a lightweight Granger causality test to identify temporal dependencies, encode domain knowledge and inferred relationships into a structural causal model (SCM), and fine-tune an LLM on aligned, expert-annotated causal explanations to ensure accuracy and actionability. Evaluated on a real-world building dataset, our system achieves significant improvements in both the factual correctness and human interpretability of causal explanations. It enables domain experts to rapidly pinpoint root causes of energy inefficiencies, thereby establishing a novel paradigm for energy diagnostics in smart buildings.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Knowledge Representation and Reasoning: Action, Change, and Causality

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Smart buildings generate vast streams of sensor and control data, but facility managers often lack clear explanations for anomalous energy usage. We propose InsightBuild, a two-stage framework that integrates causality analysis with a fine-tuned large language model (LLM) to provide human-readable, causal explanations of energy consumption patterns. First, a lightweight causal inference module applies Granger causality tests and structural causal discovery on building telemetry (e.g., temperature, HVAC settings, occupancy) drawn from Google Smart Buildings and Berkeley Office datasets. Next, an LLM, fine-tuned on aligned pairs of sensor-level causes and textual explanations, receives as input the detected causal relations and generates concise, actionable explanations. We evaluate InsightBuild on two real-world datasets (Google: 2017-2022; Berkeley: 2018-2020), using expert-annotated ground-truth causes for a held-out set of anomalies. Our results demonstrate that combining explicit causal discovery with LLM-based natural language generation yields clear, precise explanations that assist facility managers in diagnosing and mitigating energy inefficiencies.

Problem

• Explaining anomalous energy usage in smart buildings
• Providing human-readable causal explanations for energy patterns
• Diagnosing energy inefficiencies using sensor data and LLMs

Innovation

• Two-stage framework integrates causality analysis with LLM
• Lightweight causal inference module applies Granger causality tests
• LLM fine-tuned for generating actionable causal explanations