Kaya指針ハイブリッド深層学習による国別CO2排出量のleave-country-out推計：国別クラスタ推論・conformal不確実性・説明可能AIを備えた112カ国ベンチマーク
Kaya-guided hybrid deep learning for leave-country-out estimation of national CO2 emissions: a 112-country benchmark with country-clustered inference, conformal uncertainty, and explainable AI (原題)
ATUAHENE, SAMIUEL C., Boateng, Bismark, Amarfio, Hilda, Quayson, Ekow Yaleh, Agyenim, Francis Boateng
🤖 gxceed AI 要約
日本語
Kaya恒等式に基づく対数線形バックボーンとCNN-BiLSTM-Attentionの非線形残差学習を組み合わせたKG-CBAを提案し、112カ国パネルで国別除外交差検証を実施。Kaya指針は未誘導モデルより有意に精度を改善し（中央値APE 15.7%対20.8%）、リッジ回帰が全木集成・未誘導深層学習を上回った。低炭素エネルギー比率の追加で誤差が半減し、モデル高度化よりエネルギーミックスデータの有無が制約と判明。conformal区間は90%名目で81.9%カバレッジ。
English
The authors propose KG-CBA, a Kaya-guided CNN-BiLSTM-Attention model that learns only the nonlinear residual over a log-linear backbone, benchmarked on a 112-country panel under country-disjoint cross-validation. Kaya guidance significantly improves accuracy (median APE 15.7% vs 20.8%), and ridge regression outperforms all tree ensembles and unguided deep networks. Adding low-carbon energy share halves median error, showing data availability, not model sophistication, is the binding constraint; conformal intervals reach 81.9% coverage at 90% nominal.
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
パリ協定の強化透明性枠組み（ETF）やインベントリ整備が進まない国々の排出量推計に資する研究。日本企業・政策実務への直接の示唆は限定的だが、SSBJ・ISSB開示の基礎となる国別排出データの不確実性評価手法として、Scope 3やサプライチェーン排出量の国別原単位推計に応用可能な視点を提供する。
In the global GX context
Directly supports the Paris Agreement's Enhanced Transparency Framework by offering a machine-learning estimator for countries lacking inventory capacity, with honest uncertainty quantification. For global disclosure scholarship, it demonstrates that energy-mix data availability, not model complexity, constrains national emission estimation—a finding relevant to how ISSB/CSRD users assess the reliability of country-level emission factors underlying Scope 3 calculations.
👥 読者別の含意
🔬研究者:国別排出推計における機械学習の限界とデータ制約を定量化したベンチマークとして、conformal不確実性やグループ分割下の非交換性の扱いを学べる。
🏢実務担当者:サプライチェーン排出量算定に用いる国別排出原単位の不確実性を理解し、データ整備が不十分な地域の推計リスクを把握する材料になる。
🏛政策担当者:インベントリ未整備国の透明性報告支援や、排出推計の不確実性を踏まえた国際比較・政策評価の設計に示唆を与える。
📄 Abstract（原文）
Reliable national CO2 inventories underpin the Paris Agreement’s Enhanced Transparency 1national emissions modelling as leave-country-out (LOCO) estimation: predicting a country’s per-capita CO2 from eight-year trajectories of energy and socioeconomic covariates, with no access to its own emission history. Using Global Carbon Budget and Energy Institute data harmonized by Our World in Data, we build a 112-country panel (1990-2023; 3,024 windowed samples) and benchmark thirteen models under country-disjoint five-fold cross-validation. The proposed Kaya-guided CNN-BiLSTM-Attention network (KG-CBA) couples a log-linear backbone motivated by the Kaya identity to a convolutional-recurrent-attention branch that learns only the nonlinear residual, tuned by particle swarm optimization. Because the sample contains 27 repeated years per country, all inference is country-clustered. Kaya guidance significantly improves the unguided architecture (median APE 15.7% vs 20.8%, p = 0.002), and ridge regression significantly outperforms every tree ensemble and every unguided deep network; but the best ensemble’s apparent edge over ridge is not significant (15.1% vs 16.3%, p = 0.22), where observation-level testing would report p < 10-9. Error is highly concentrated: six structurally atypical energy systems contribute 79-84% of total squared error. Adding the low-carbon energy share on a 76-country sub-panel halves median error (7.35% vs 13.72% like-for-like), showing that energy-mix data availability, not model sophistication, is the binding constraint. Removing all CO2-derived covariates costs under one percentage point, so the estimator suits countries with no inventory at all. Split-conformal Monte-Carlo-dropout intervals reach 81.9% coverage at a 90% nominal level, the shortfall traced to non-exchangeability under grouped splits. In conventional temporal forecasting, learners beat naive persistence only marginally and inconsistently across evaluation windows. Data and code are open.
🔗 Provenance — このレコードを発見したソース
- EarthArXiv https://eartharxiv.org/repository/object/14906/download/25910/first seen 2026-09-10 04:13:26 · last seen 2026-09-21 04:15:59
🔔 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。