Skip to content

HKUST-KnowComp/ASER

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ASER (Activities, States, Events, and their Relations)

logo

ASER is a large-scale weighted eventuality knowledge graph, including actions, states, events, and their relations.

The eventualities (i.e., nodes of ASER) are extracted using selected dependency patterns. The edges are based on discourse relations (e.g., Result) in discourse analysis.

Besides, conceptualized eventualities in a more abstract level and their relations are also conducted to generalize the knowledge.

In total, ASER (full) contains 438 million eventualities and 648 million edges between eventualities; ASER (core) contains 53 million eventualities and 52 million edges between eventualities.

With the help of Probase (now called Microsoft Concept Graph), ASER (concept) contains 15 million conceptualized eventualities and 224 million edges between conceptualied eventualities. We also provide a copy of Probase download from MSRA's official website. All licenses are subject to MSRA's original release.

The homepage of the project and data is https://hkust-knowcomp.github.io/ASER.

The online demo is coming soon.

  • ASER 2.1 (dev): using original text tokens as eventualities (set use_lemma=False when using extractors) and checking the completeness via the dependency parser. [code branch] [data]

  • ASER 2.0 (AIJ"2022): ASER: Towards Large-scale Commonsense Knowledge Acquisition via Higher-order Selectional Preference over Eventualities. [pdf] [code branch] [data]

  • ASER 1.0 (WWW"2020): ASER: A Large-scale Eventuality Knowledge Graph. [pdf] [code branch] [data]

Quick Start

Please refer to the get_started.ipynb or documentation to become familiar with ASER and its construction pipeline.

References

@article{ZhangLPKOFS22,
  author    = {Hongming Zhang and
               Xin Liu and
               Haojie Pan and
               Haowen Ke and
               Jiefu Ou and
               Tianqing Fang and
               Yangqiu Song},
  title     = {{ASER:} Towards Large-scale Commonsense Knowledge Acquisition via Higher-order Selectional Preference over Eventualities},
  journal   = {Artificial Intelligence},
  volume    = {309},
  pages     = {103740},
  year      = {2022},
}

@inproceedings{ZhangLPSL20,
  author    = {Hongming Zhang and
               Xin Liu and
               Haojie Pan and
               Yangqiu Song and
               Cane Wing{-}Ki Leung},
  title     = {{ASER:} {A} Large-scale Eventuality Knowledge Graph},
  booktitle = {WWW},
  pages     = {201--211},
  year      = {2020}
}

About

ASER (Activities, States, Events, and their Relations): a large-scale weighted eventuality knowledge graph.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages