Enhanced ESG Report Analysis using Modified Multimodal RAG
Bibliographic record
Abstract
Environmental, Social and Governance (ESG) criteria have emerged as fundamental pillars for evaluating corporate sustainability, ethical stewardship and governance practices, increasingly guiding investor strategies and public trust. Amid escalating climate challenges, social inequalities and regulatory pressures, transparent and comprehensive ESG reporting has become imperative for enterprises to ensure compliance, attract responsible investment and build long-term resilience. However, ESG reports are inherently complex, encompassing unstructured narratives alongside dense visual data—such as graphs, flowcharts and images—that carry critical contextual and quantitative insights. Traditional text-based language models and document retrieval systems do not address these multimodal complexities, often overlooking or misinterpreting visual components essential for a nuanced assessment of ESG. To bridge this gap, this research introduces an enterprise-grade multimodal Retrieval-Augmented Generation (RAG) framework that goes beyond conventional text embeddings by integrating advanced image, chart and layout processing capabilities. This multimodal approach not only enables dynamic retrieval of external domain knowledge but also ensures contextually precise, holistic insights into ESG disclosures, empowering enterprises with an intelligent, scalable tool to navigate the multifaceted landscape of modern ESG reporting.The empirical evaluation demonstrates a measurable improvement of 6.5% in mean average precision over the strongest prior baseline, validating the framework’s ability to deliver quantifiable enhancements in ESG report analysis.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.011 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.004 | 0.002 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.003 | 0.002 |
| Open science | 0.001 | 0.003 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.009 | 0.005 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".