What is Ascent++?
Ascent++ KB is a commonsense knowledge base (CSKB) constructed from the C4 crawl using the Ascent++ pipeline. It consists of 2 million CSK assertions about 10K popular concepts, presented in the established ConceptNet schema with 19 predicates (e.g., AtLocation, CapableOf, HasProperty, etc.). As of 2022, it presents the highest-quality automated CSKB, both in terms of precision, and in terms of ranked recall.
Ascent++, a successor of Ascent, is a pipeline for automatically collecting, extracting and consolidating commonsense knowledge (CSK) from any English text corpus. Ascent++ is capable of extracting facet-enriched assertions, overcoming the common limitations of the triple-based knowledge model in traditional knowledge bases (KBs). Ascent++ also captures composite concepts with subgroups and related aspects, supplying even more expressiveness to CSK assertions.
This web interface allows you to browse the CSK assertions in the Ascent++ KB.
Examples
To explore what our CSKB captures, try out the following concepts:
Download
You can download 2 million CSK assertions in our KB here: ascentpp.csv.tar.gz.
Code
Code is stored in this Github repository.
Citation
Main publication
@ARTICLE{ascentpp, author={Nguyen, Tuan-Phong and Razniewski, Simon and Romero, Julien and Weikum, Gerhard}, journal={IEEE Transactions on Knowledge and Data Engineering}, title={Refined Commonsense Knowledge from Large-Scale Web Contents}, year={2022}, doi={10.1109/TKDE.2022.3206505} }
Related publications
@inproceedings{ascent, author = {Nguyen, Tuan-Phong and Razniewski, Simon and Weikum, Gerhard}, title = {Advanced Semantics for Commonsense Knowledge Extraction}, year = {2021}, isbn = {9781450383127}, publisher = {Association for Computing Machinery}, address = {New York, NY, USA}, doi = {10.1145/3442381.3449827}, booktitle = {Proceedings of the Web Conference 2021}, pages = {2636–2647}, numpages = {12}, location = {Ljubljana, Slovenia}, series = {WWW '21} }
@inproceedings{uncommonsense, author = {Arnaout, Hiba and Razniewski, Simon and Weikum, Gerhard and Pan, Jeff Z.}, title = {UnCommonSense: Informative Negative Knowledge about Everyday Concepts}, year = {2022}, isbn = {9781450392365}, publisher = {Association for Computing Machinery}, address = {New York, NY, USA}, doi = {10.1145/3511808.3557484}, booktitle = {Proceedings of the 31st ACM International Conference on Information & Knowledge Management}, pages = {37–46}, numpages = {10}, keywords = {knowledge base, commonsense, negation}, location = {Atlanta, GA, USA}, series = {CIKM '22} }
@inproceedings{quasimodo, author = {Romero, Julien and Razniewski, Simon and Pal, Koninika and Z. Pan, Jeff and Sakhadeo, Archit and Weikum, Gerhard}, title = {Commonsense Properties from Query Logs and Question Answering Forums}, year = {2019}, isbn = {9781450369763}, publisher = {Association for Computing Machinery}, address = {New York, NY, USA}, doi = {10.1145/3357384.3357955}, booktitle = {Proceedings of the 28th ACM International Conference on Information and Knowledge Management}, pages = {1411–1420}, numpages = {10}, keywords = {common-sense knowledge acquisition, web mining, information extraction}, location = {Beijing, China}, series = {CIKM '19} }