Summary of Statements: Universal Information Extraction From Tables with Large Language Models For Esg Kpis, by Lokesh Mishra et al.
Statements: Universal Information Extraction from Tables with Large Language Models for ESG KPIs
by Lokesh Mishra, Sohayl Dhibi, Yusik Kim, Cesar Berrospi Ramis, Shubham Gupta, Michele Dolfi, Peter Staar
First submitted to arxiv on: 27 Jun 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary A novel deep-learning approach, Statements, is proposed for extracting quantitative facts and related information from Environment, Social, and Governance (ESG) reports’ tables. This is a crucial step in assessing an organization’s performance on climate change, greenhouse gas emissions, water consumption, waste management, human rights, diversity, and policies. The proposed method involves translating tables to statements as a new supervised task, utilizing SemTabNet, a dataset of over 100K annotated tables. A family of T5-based Statement Extraction Models is investigated, with the best model generating statements that are 82% similar to ground-truth. The advantages of this approach are demonstrated by applying the model to over 2700 ESG report tables, enabling exploratory data analysis on large collections. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Environment, Social, and Governance (ESG) reports help track an organization’s progress on important issues like climate change and human rights. But extracting useful information from these reports is hard because table structures are different and the content can be confusing. To fix this, scientists created a new way to turn tables into easy-to-understand statements. They developed a special kind of artificial intelligence that can learn to extract these statements from large collections of ESG report data. This new approach could help people quickly analyze big sets of information and make better decisions. |
Keywords
» Artificial intelligence » Deep learning » Supervised » T5
