Patent Portfolio Records Question Answering Dataset

#document question answering #evidence-based reasoning #patent information extraction #patent portfolio management #intellectual property research #patent information retrieval

5000
0.4GB
JSON
CATL
MOBIUSI INC

Updated:2026-10-18

AI Analysis & Value Prop

This dataset provides closed-domain question-answering examples grounded in patent applications, granted patent texts, and portfolio records, with expert answers and supporting documentary evidence. Its topics include patent families, priority claims, claims, legal status, and geographic coverage; structured portfolio facts and scope notes help trace answer support and identify information that is absent or cannot be confirmed from the documents. It is designed for domain-specific fine-tuning, closed-document QA evaluation, and queries about portfolio composition and recorded facts.

Dataset Insights

Sample Examples

{}

Technical Specifications

Field	Type	Description
file_name	string	File name
answer	string	An accurate, clear answer to the user question, limited to facts recorded in the source documents.
evidence	array	Source document references, supporting excerpts, and the corresponding fact categories.
question	string	A patent portfolio management question that must be answered using the supplied source documents.
scope_note	string	A note describing the document scope behind the answer, missing information, or matters that cannot be confirmed from the available documents. Omit when no note is needed.
portfolio_facts	object	Patent portfolio facts derived from the source documents. Include only information explicitly supported by the documents; omit anything that cannot be confirmed.
source_documents	array	Patent application texts, granted patent texts, or portfolio records used to support the answer. Each record corresponds to one document.

Compliance Statement

Authorization Type	Proprietary - Commercial AI Training License (No Redistribution)
Commercial Use	Requires exclusive subscription or authorization contract (monthly or per-invocation charging)
Privacy and Anonymization	No PII, no real company names, simulated scenarios follow industry standards
Compliance System	Compliant with China's Data Security Law / EU GDPR / supports enterprise data access logs

Frequently Asked Questions

What does the dataset contain?: Each example is grounded in a patent application, granted patent text, or portfolio record. It includes source documents, a user question, an expert answer, and documentary evidence, and may also provide structured portfolio facts and a scope note. Topics include patent families, priority claims, claims, legal status, and geographic coverage.

How are answers linked to their source documents?: Answers are designed to rely only on facts explicitly recorded in the supplied documents. Evidence fields provide supporting citations or excerpts and identify the relevant fact category. A scope note can flag matters that the documents do not cover or cannot confirm, supporting traceable closed-domain QA training and evaluation.

Which model development and business use cases does it support?: The dataset can support domain-specific supervised fine-tuning for patent portfolio management, closed-document QA evaluation, and queries about patent assets, portfolio records, and related legal status facts. For each use case, ensure that the available document scope matches the target task.

How can the dataset be obtained, and what are the licensing terms?: Delivery format, access method, update schedule, and licensing scope should be confirmed in the data provider's product terms. Before purchase or integration, verify that the license covers the intended model training, evaluation, and commercial deployment, and observe the applicable rights governing source documents and related data.

Can't find the data you need?

Post a request and let data providers reach out to you.

Get this Dataset

Verified for Enterprise Use

Cite this Work

@dataset{Mobiusi2026,
  title={Patent Portfolio Records Question Answering Dataset},
  author={MOBIUSI INC},
  year={2026},
  url={https://www.mobiusi.com/datasets/164bf0a8c471edab4c8398961de47d1a},
  urldate={2026-10-12},
  keywords={patent portfolio QA dataset, patent records data, patent document QA, patent families, patent legal status, portfolio management, legal domain SFT, closed-domain QA},
  version={1.0}
}

Using this in research? Please cite us.

placeholder

Products

Scene

Domain

Modality

Task