A dataset of publications and patents on quantum error correction (1995-2025)

This article presents a harmonized corpus of scholarly publications and patents concerning quantum error correction, the encoding techniques that protect quantum information against noise and that underpin the pursuit of fault-tolerant quantum computation. The corpus spans 1 January 1995 to 31 December 2025 and contains 16,274 records: 14,385 scholarly publications harvested from OpenAlex and enriched from Crossref, and 1889 United States patent publications retrieved from Google Patents Public Data. Publications were collected by matching twenty scope phrases against titles and abstracts, intersected with a quantum-computing topic to suppress unrelated matches; patents were collected under the Cooperative Patent Classification class for quantum error correction. Both record types were mapped onto a common eighteen-field schema, deduplicated within type, and assigned to one of three retrieval tiers according to the strength of the topical match. Each record carries subfield tags and flags marking machine-learning decoder records and adjacent-topic mentions. The precision of the core tier was assessed by manual annotation of a stratified sample, and inter-annotator agreement was measured. The corpus, its variable dictionary, provenance metadata, validation samples, and the complete construction notebook are openly available, supporting reuse in scientometrics, patent analysis, and text mining of the field.