Alan Akbik and Thilo Michael
The Weltmodell: A Data-Driven Commonsense Knowledge Base
Abstract:
We present the WELTMODELL, a commonsense knowledge base that was automatically generated from aggregated dependency parse fragments gathered from over 3.5 million English language books. Using techniques from open-domain Information Extraction (IE) and distributional semantics, we arrive at close to ten million distinct N-ary commonsense facts as well as a range of measures of association and similarity. We discuss future directions of mining and working with commonsense knowledge, as well as limitations of unsupervised approaches. The Weltmodell can be browsed online at: www.textmining.tu-berlin.de/weltmodell/
Download: Link
