Archive Notice: This path historically hosted open-source prototypes from the Information Retrieval Facility (IRF) before the organization wound down operations in 2012. This domain now hosts QuizLevelUp.
IRF Prototypes & MAREC Dataset
This directory historically served as the primary download interface and evaluation hub for the MAREC (Matrixware Research Collection) dataset.
What was the MAREC Prototype?
It was a standardized, massive multilingual patent corpus featuring over 19 million patent documents (spanning 1976–2008) normalized to a highly specific, uniform XML schema. Researchers utilized this data to design, train, and benchmark advanced semantic search systems, machine translation tools, and natural language processing (NLP) models.
Active Resources for Researchers
- View historical captures of this page on the Internet Archive Wayback Machine .
- Review the original structural definitions on the official MAREC Wikipedia Page and download the research corpus from TU Wien .
- Download modern open bulk text datasets from the official USPTO Open Data Portal .
- Access large-scale technical patent documentation directly via the European Patent Office (EPO) Bulk Data Sets portal.
Building Data Pipelines or Software?
While the old academic data repository is permanently offline, this domain now hosts QuizLevelUp—a platform built explicitly for data engineers and developers to test their technical skills in Data Architecture, AI Systems, Cloud Engineering, and Cybersecurity.
Explore QuizLevelUp Trivia Tracks 🚀