Federated Learning-Enabled Building Stock Modeling for Privacy-Preserving Embodied Carbon Benchmarking in Residential Construction
連合学習によるプライバシー保護型建築ストックモデリング：住宅建設における埋込炭素ベンチマーク (AI 翻訳)
N. Albelwi
🤖 gxceed AI 要約
日本語
本研究は、連合学習に基づく建築ストックモデリングシステム「FedCarbon」を提案する。データを一箇所に集約せずに建設関係者が協調して埋込炭素を評価可能とし、差分プライバシーと注意機構により精度とプライバシーを両立。768件の住宅構成データと2340件の欧州データで検証し、R2=94.2%、MAE=21.4 kgCO2e/m2を達成した。
English
This paper introduces FedCarbon, a federated learning-based building stock modeling system that enables collaborative embodied carbon benchmarking without central data aggregation. Using hierarchical federated aggregation with attention-based client weighting and adaptive differential privacy, it achieves 94.2% R2 and MAE of 21.4 kgCO2e/m2 on two datasets (UCI Energy Efficiency and European buildings database), while ensuring data sovereignty and reducing communication overhead by 82.6%.
Unofficial AI-generated summary based on the public title and abstract. Not an official translation.
📝 gxceed 編集解説 — Why this matters
日本のGX文脈において
日本の建築業界では、SSBJ対応やサプライチェーン排出量算定が求められる中、競合間でのデータ共有は困難。FedCarbonのようなプライバシー保護型連合学習は、建設会社や材料メーカーが協調してベンチマークを構築し、脱炭素を加速する上で重要な技術となる。
In the global GX context
Globally, embodied carbon benchmarking faces data sharing barriers due to privacy and competitive concerns. FedCarbon addresses this by enabling collaborative model training without centralizing data, which is crucial for industry-wide low-carbon strategies. The approach aligns with TCFD and ISSB disclosure trends by providing a scalable, privacy-compliant carbon assessment method applicable across jurisdictions.
👥 読者別の含意
🔬研究者:This work demonstrates a novel federated learning architecture with differential privacy for carbon accounting, offering a template for privacy-preserving collaborative modeling in sustainability.
🏢実務担当者:Construction firms and material suppliers can adopt FedCarbon to benchmark embodied carbon across projects without exposing proprietary data, aiding in Scope 3 disclosure and competitive analysis.
🏛政策担当者:Policymakers can leverage such privacy-preserving frameworks to mandate industry-wide carbon benchmarking without requiring centralized data collection, balancing transparency with data sovereignty.
📄 Abstract（原文）
Benchmarking embodied carbon in residential building stock accurately would involve a high volume of data sharing and would pose serious privacy and competitive issues among building construction stakeholders. This study introduces a new federated learning-based building stock modeling system (FedCarbon) that can allow embodied carbon to be evaluated collaboratively without data aggregation at a central place. The architecture proposed enables construction firms, cities, and providers of construction materials to collectively train predictive models at the same time as data sovereignty is achieved via a hierarchical federated aggregation mechanism with attention-based client weighting. A differentiated privacy scheme that is adaptively calibrated on noise guarantees the privacy of individual projects and allows for statistically significant benchmarking based on heterogeneous building portfolios. The framework also includes a gradient compression scheme based on momentum, which incurs an 82.6% reduction in communication overhead over traditional federated averaging-based methods and still maintains model convergence. The effectiveness of the approach is demonstrated with the help of comprehensive validation with the UCI Energy Efficiency Dataset, which includes 768 residential building configurations, and the Embodied Carbon in European Buildings Database, which includes 2340 residential units in 12 European jurisdictions. It has been experimentally shown that FedCarbon has a 94.2% prediction accuracy (R2) on embodied carbon intensity, with a mean absolute error of 21.4 kgCO2e/m2, and that (ε, δ) differential privacy can be guaranteed with ε = 1.0 and −δ = 10−5. This structure opens up building stock knowledge and hastens industry-wide implementation of low-carbon building strategies.
🔗 Provenance — このレコードを発見したソース
- semanticscholar https://doi.org/10.3390/buildings16051029first seen 2026-06-29 06:29:59
🔔 こうした論文の新着を逃したくない方は キーワードアラート に登録（無料・3キーワードまで）。
gxceed は公開メタデータに基づく研究支援データセットです。要約・翻訳・解説は AI 支援で生成されています。 最終的な解釈・検証は利用者が原典資料に基づいて行うことを前提とします。