aiqbench analysis
This paper introduces HiFi-KPI, a large-scale dataset designed to improve the extraction of Key Performance Indicators from financial filings using machine learning. By addressing the limitations of complex iXBRL taxonomies, the study evaluates encoder-based models and LLMs on tasks like classification and structured extraction, highlighting specific challenges with date-related errors. The release of this resource aims to enhance automated analysis of earnings reports for stakeholders seeking short-term returns based on accurate financial data tagging.
HiFi-KPI: A Dataset for Hierarchical KPI Extraction from Earnings Filings
arXiv:2502.15411v5 Announce Type: replace Abstract: Accurate tagging of earnings reports can yield significant short-term returns for stakeholders. The machine-readable inline eXtensible Business Reporting Language (iXBRL) is mandated for public financial filings. Yet, its complex, fine-grained taxonomy limits the cross-company transferability of tagged Key Performance Indicators (KPIs). To address this, we introduce the Hierarchical Financial Key Performance Indicator (HiFi-KPI) dataset, a large-scale corpus of 1.65M paragraphs and 198k unique, hierarchically organized labels linked to iXBRL taxonomies. HiFi-KPI supports multiple tasks and we evaluate three: KPI classification, KPI extraction, and structured KPI extraction. For rapid evaluation, we also release HiFi-KPI-Lite, a manually curated 8K paragraph subset. Baselines on HiFi-KPI-Lite show that encoder-based models achieve over 0.906 macro-F1 on classification, while Large Language Models (LLMs) reach 0.440 F1 on structured extraction. Finally, a qualitative analysis reveals that extraction errors primarily relate to dates. We open-source all code and data at https://github.com/aaunlp/HiFi-KPI.
Abstract:Accurate tagging of earnings reports can yield significant short-term returns for stakeholders. The machine-readable inline eXtensible Business Reporting Language (iXBRL) is mandated for public financial filings. Yet, its complex, fine-grained taxonomy limits the cross-company transferability of tagged Key Performance Indicators (KPIs). To address this, we introduce the Hierarchical Financial Key Performance Indicator (HiFi-KPI) dataset, a large-scale corpus of 1.65M paragraphs and 198k unique, hierarchically organized labels linked to iXBRL taxonomies. HiFi-KPI supports multiple tasks and we evaluate three: KPI classification, KPI extraction, and structured KPI extraction. For rapid evaluation, we also release HiFi-KPI-Lite, a manually curated 8K paragraph subset. Baselines on HiFi-KPI-Lite show that encoder-based models achieve over 0.906 macro-F1 on classification, while Large Language Models (LLMs) reach 0.440 F1 on structured extraction. Finally, a qualitative analysis reveals that extraction errors primarily relate to dates. We open-source all code and data at this https URL.
Camera-ready. Accepted at LREC 2026 (main conference)
Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
arXiv:2502.15411 [cs.CL]
(or arXiv:2502.15411v5 [cs.CL] for this version)
Proceedings of the Fifteenth Language Resources and Evaluation Conference (LREC 2026), pages 441-455, Palma, Mallorca, Spain. ELRA, 2026
https://doi.org/10.63317/2nbsp7zzfb3g DOI(s) linking to related resources
From: Rasmus T. Aavang [view email] [v1] Fri, 21 Feb 2025 12:19:08 UTC (1,000 KB) [v2] Mon, 24 Feb 2025 14:45:27 UTC (1,003 KB) [v3] Thu, 19 Mar 2026 10:06:48 UTC (2,188 KB) [v4] Mon, 1 Jun 2026 13:37:16 UTC (2,188 KB) [v5] Thu, 20 Aug 2026 12:06:58 UTC (2,188 KB)
Full text under the publisher's open licence · 242 words
Related coverage
Article: Governing AI in the Cloud: A Practical Guide for Architects
InfoQ · Jun 15, 2026
Configurable Semantic Chunking for Biomedical Information Extraction in Retrieval-Augmented Generation
arXiv · Sep 1, 2026
LEDGER: A Long-Context Benchmark of Corporate Annual Reports for Grounded Financial Retrieval and Extraction
arXiv · Jun 12, 2026
Build an end-to-end RFI questionnaire workflow using Amazon Quick Automate
AWS · Sep 10, 2026