GX & Decarbonization Research

Kingsuk Majumdar, Sohini Ghosh

This study applies Random Forest, XGBoost, and SVM to a 65-country, five-sector EU ETS trade panel (2000–2018) to predict carbon leakage. XGBoost achieves R²=0.965 for embodied CO2 regression and ROC-AUC=0.765 for leakage classification, su…

Murali Krishna Pasupuleti

This monograph proposes a research architecture linking financial statements, management commentary, sustainability metrics, climate-risk evidence and operational telemetry via multimodal AI. It frames reporting quality as a joint function …

Qiao-Rong Yang

A critical review of 15 studies at the intersection of AI, carbon accounting, and green accounting. It identifies three technical pathways: data-centric systems, ontology-driven hybrids for SMEs, and real-time tagging platforms. The paper h…

Anonymous

Replication deposit for a firm-level study of CSRD climate disclosure quality before and after the 2026 Omnibus scope reduction. Covers 188 EU firms and 376 reports coded on a ten-dimension instrument using a frozen LLM prompt. Includes the…

Adeboye Oyegunle, Olaf Weber

This study uses AI and NLP to examine how well 16 major global climate risk banking regulations align with 83 peer-reviewed articles (2017–2024). It finds only moderate alignment (74.2% cosine similarity), showing regulators partially incor…

(著者不明)

This study examines how CEO narcissism affects the quality of TCFD-aligned climate-related financial disclosure among S&P 1500 firms. Using generative NLP measures across the four TCFD pillars, it finds narcissistic CEOs produce weaker, les…

Naveen Virmani, Srikant Gupta, Koppiahraj Karuppiah +1

This paper examines federated learning as a means to enhance the scalability and credibility of ESG reporting, identifying and prioritizing adoption challenges. Through literature review and expert panels, 18 validated challenges were analy…

Xunzhang Pan, Tianming Shao, Jianxiao Wang +5

This paper uses LLM agents to generate and evaluate global mitigation pathways for the 1.5°C target, integrating Sustainable Development Goals (SDGs). It sits at the intersection of AI and climate policy, offering a novel approach that comb…

Marcus Becker, Timo Zimmermann, Judith Maria Beermann

This study applies a GRI-oriented RAG workflow to eight years of Borussia Dortmund's sustainability reports, finding moderate and fluctuating evidence quality rather than steady improvement. It demonstrates RAG's utility in structuring disc…

Jin Yan, Yang Zhao, Xue Gao +3

Combining machine learning with causal inference, this study detects low-carbon technology innovation take-offs across 18 countries (2001-2022) and evaluates over 1,600 policy interventions. It finds that market-based instruments (especiall…

Rui Lu

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.

Filiz Mızrak, Umut Elbir

This study develops a reliability-aware Bayesian framework to assess readiness for digital-twin-enabled smart-energy systems across 20 major international airports. Using 15 capability indicators in four domains plus an auxiliary transparen…

Baptiste Rigaux, Mark Vergouwen

DETeCT proposes a data-efficient method for classifying low-carbon technologies into multiple categories. It achieves high classification performance with limited data, aiming to automate technology classification. This can support corporat…

Daichen Guo, Sanglin Zhao

This study empirically examines the impact of corporate digital transformation (DX) on carbon emissions and the mediating role of ESG performance, using Chinese A-share listed companies from 2010 to 2023. DX reduces carbon emissions by impr…

Yuanyuan Ma, Siyu Peng, Yun Yu

This study develops a hybrid forecasting framework integrating CEEMDAN, sample entropy reconstruction, and coefficient of variation ensemble to capture multi-scale nonlinear carbon price dynamics. SHAP and TVP-SV-VAR reveal scale-specific a…

Sohail M.

This paper empirically examines how AI adoption affects ESG performance in the German banking sector, focusing on the mediating role of green finance. It suggests that AI can enhance banks' ESG assessments and promote green lending.

Xuelong Zhang, Xiaoling He, Mei Li +3

Using panel data of Chinese listed energy firms (2010-2023), this study shows AI adoption significantly reduces carbon emission intensity. Executive digital background strengthens the effect, and low-carbon city pilots provide complementary…

Saad Khan, Ahmed A. Aboushady, Firdous Nazir +1

This study proposes a privacy-preserving framework to detect, classify, and localize unregistered behind-the-meter low-carbon technologies (PV, EVs, heat pumps) in low-voltage networks. Using GDPR-compliant measurements, it combines LSTM au…

Aurélien Saussay, Misato Sato, Francesco Vona

Using U.S. online job vacancy data, this paper develops a skill-based NLP method to identify low-carbon jobs within occupations. Low-carbon job creation is prevalent in low-skilled occupations but requires more complex skills, yielding a mo…

Issam Benhayoun

This study analyzes factors determining ISSB adoption intention among 335 Moroccan accounting professionals. It demonstrates that institutional pressures influence adoption intention only through perceived usefulness and ease of use. Using …