The document describes the 2014 Linked Data for Information Extraction Challenge, focusing on the task of creating a system to extract structured information from HTML websites using microformats. It details the creation of a training dataset from annotated web pages, the methodology for extracting RDF triples using the Any23 tool, and the evaluation process for the submissions with a goal to improve upon existing baselines. Additionally, it outlines plans for future challenges, highlighting the need for more diverse classes and realistic settings.