GreenPassport: クロスボーダーAI推論のリクエストレベル炭素会計
GreenPassport: Request-Level Carbon Accounting for Cross-Border AI Inference (原題)
Rui Lu
🤖 gxceed AI 要約
日本語
AI推論の炭素排出をリクエスト単位で算定する手法を提案。サービス境界、経路、電力源等を考慮し、公開データで実装。EcoLogitsと比較し誤差を大幅削減。
English
Proposes request-level carbon accounting for AI inference, considering service boundaries, routes, and electricity sources. Public-data implementation reduces median absolute percentage error by 56.3% vs EcoLogits.
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
日本ではSSBJ開示やサプライチェーン排出算定が進む中、AI利用のScope2・3算定精度向上に寄与。クラウド利用の排出可視化は投資家対応にも有用。
In the global GX context
Addresses a gap in AI-related carbon accounting, relevant to ISSB/CSRD disclosure where digital services' indirect emissions are increasingly scrutinized. Offers a methodology for more accurate reporting.
👥 読者別の含意
🔬研究者:AI推論の炭素算定における境界設定と精度向上の方法論を提供。
🏢実務担当者:クラウド利用の排出算定精度を高め、開示対応に活用可能。
🏛政策担当者:デジタル分野の排出算定基準策定に示唆を与える。
📄 Abstract（原文）
AI inference often crosses regional boundaries as prompts travel to remote data centers and generated tokens return to users. Regional averages cannot represent the resulting differences in serving hardware, electricity, and network delivery. Request-level accounting needs a common boundary for the service, serving site, route, local comparator, uncertainty, and data provenance.GreenPassport Carbon Accounting (GPCA) associates these inputs with each request. It estimates serving and route carbon, then selects a reporting level from the available documentation. Our public-data implementation covers data-center instances, accelerators, model families, electricity mixes, routes, and cloud-region carbon intensity. Against six accounting baselines and four energy-prediction baselines, GPCA reduced median absolute percentage error by 56.3\% and median absolute error by 15.5\% relative to EcoLogits under the aligned accelerator-energy boundary. It produced zero rule overstatement in the deterministic conformance tests. In the buyer case, the clean-electricity CN-West scenario produced $0.0148$ gCO$_2$e /request, 88\% below the local service at $0.1220$ gCO$_2$e /request.
🔗 Provenance — このレコードを発見したソース
- arXiv https://arxiv.org/abs/2609.06784first seen 2026-09-09 03:53:22 · last seen 2026-09-10 04:10:38
🔔 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。