Dataset
AIPD_nlp_granted_claims
Dataset card entry for "AIPD_nlp_granted_claims" with additional details required.
Data preview
Dataset card
AIPD_nlp_granted_claims is a Hugging Face dataset of U.S. granted patent claims reformatted as instruction, input, and output triples for natural language processing research. It is distributed in Parquet and has accumulated modest community downloads.
What this dataset contains
The dataset packages granted U.S. patent claims as supervised learning examples, pairing claim-rewriting instructions with the original and generated claim text. Per the published schema, each record links a patent number to a parent and dependent claim, and the total is just over one million rows split across train, test, and valid sets. The publisher is listed simply as "patent," and the license is not stated, so reuse rights are unclear.
Structure and fields
Six columns define each row. The identifiers patent_num (int64), claim_num1 (int64), and claim_num2 (int64) locate the source claim and its counterpart. The text fields are instruction (string), input (string), and output (string); for example, the instruction may direct the model to generate a parent or child claim, the input holds the dependent claim text, and the output returns the rewritten parent or child claim. Data is delivered in Parquet and is compatible with the datasets, dask, mlcroissant, and polars libraries.
How analysts use it
The dataset is primarily a training corpus for NLP models that draft or rewrite patent claims, which makes it a public-records asset rather than a direct market signal. Investment teams tracking patent-heavy sectors could use it to monitor technology descriptions at scale, build classifiers over claim language, or benchmark generative models relevant to legal-tech and intellectual-property software vendors. The unspecified license and modest download counts suggest it is more useful for research prototypes than for production investment pipelines.
Documentation & references 1
Related datasets
Browse moreParsed eligibility criteria from 13,229 ClinicalTrials.gov studies representing the candidate pool gathered by two first-stage retrievers during TREC Clinical Trials 2021–2023, formatted as typed entity-relation graphs for reranking models.
The clinical-trials-patient-graphs dataset is a small tabular collection of patient-level clinical records distributed in Parquet format by the 2001jdev publisher. It organizes entities and relations extracted from trial data into structured rows spanning 2021 through 2023.
This dataset is a parsed snapshot of ClinicalTrials.gov records, distributed in Parquet format on the Hugging Face Hub under an unspecified license. It contains roughly 52,000 trial entries drawn from public registry filings.
clinical-trials-trec-qrels is a tabular relevance-judgment file distributed by the 2001jdev user on the Hugging Face Hub. It maps clinical-trial topic identifiers to NCT registry IDs with graded relevance scores.
The clinical-trials-trec-topics dataset is a small tabular collection of clinical-trial topic records distributed in Parquet format on the Hugging Face Hub. It appears to be derived from TREC clinical-trial retrieval topics, with rows partitioned by year.
SEC EDGAR 10-K & 10-Q financial statements for 2,600+ NASDAQ companies.
Discussion
Questions or observations about this dataset? Join the conversation on X.
No comments yet. Start the conversation on X.