b'@online{Nguyen_2011.00905,'b'\nTITLE = {Advanced Semantics for Commonsense Knowledge Extraction},\nAUTHOR = {Nguyen, Tuan-Phong and Razniewski, Simon and Weikum, Gerhard},\nLANGUAGE = {eng},\nURL = {https://arxiv.org/abs/2011.00905},\nEPRINT = {2011.00905},\nEPRINTTYPE = {arXiv},\nYEAR = {2020},\nABSTRACT = {Commonsense knowledge (CSK) about concepts and their properties is useful for<br>AI applications such as robust chatbots. Prior works like ConceptNet, TupleKB<br>and others compiled large CSK collections, but are restricted in their<br>expressiveness to subject-predicate-object (SPO) triples with simple concepts<br>for S and monolithic strings for P and O. Also, these projects have either<br>prioritized precision or recall, but hardly reconcile these complementary<br>goals. This paper presents a methodology, called Ascent, to automatically build<br>a large-scale knowledge base (KB) of CSK assertions, with advanced<br>expressiveness and both better precision and recall than prior works. Ascent<br>goes beyond triples by capturing composite concepts with subgroups and aspects,<br>and by refining assertions with semantic facets. The latter are important to<br>express temporal and spatial validity of assertions and further qualifiers.<br>Ascent combines open information extraction with judicious cleaning using<br>language models. Intrinsic evaluation shows the superior size and quality of<br>the Ascent KB, and an extrinsic evaluation for QA-support tasks underlines the<br>benefits of Ascent.<br>},\nJOURNAL = {WWW 2021},\n}\n'