CSKG: The CommonSense Knowledge Graph

被引:44
作者
Ilievski, Filip [1 ]
Szekely, Pedro [1 ]
Zhang, Bin [1 ]
机构
[1] Univ Southern Calif, Informat Sci Inst, Los Angeles, CA 90007 USA
来源
SEMANTIC WEB, ESWC 2021 | 2021年 / 12731卷
关键词
Commonsense knowledge; Knowledge graph; Embeddings;
D O I
10.1007/978-3-030-77385-4_41
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Sources of commonsense knowledge support applications in natural language understanding, computer vision, and knowledge graphs. Given their complementarity, their integration is desired. Yet, their different foci, modeling approaches, and sparse overlap make integration difficult. In this paper, we consolidate commonsense knowledge by following five principles, which we apply to combine seven key sources into a first integrated CommonSense Knowledge Graph (CSKG). We analyze CSKG and its various text and graph embeddings, showing that CSKG is well-connected and that its embeddings provide a useful entry point to the graph. We demonstrate how CSKG can provide evidence for generalizable downstream reasoning and for pre-training of language models. CSKG and all its embeddings are made publicly available to support further research on commonsense knowledge integration and reasoning.
引用
收藏
页码:680 / 696
页数:17
相关论文
共 38 条
  • [31] WebChild 2.0: Fine-Grained Commonsense Knowledge Distillation
    Tandon, Niket
    de Melo, Gerard
    Weikum, Gerhard
    [J]. PROCEEDINGS OF THE 55TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2017): SYSTEM DEMONSTRATIONS, 2017, : 115 - 120
  • [32] Trouillon T, 2016, PR MACH LEARN RES, V48
  • [33] Trumbo D., 2006, THESIS U COLORADO BO
  • [34] Wikidata: A Free Collaborative Knowledgebase
    Vrandecic, Denny
    Kroetzsch, Markus
    [J]. COMMUNICATIONS OF THE ACM, 2014, 57 (10) : 78 - 85
  • [35] Yang Bishan, 2014, 3 INT C LEARNING REP
  • [36] Yang ZL, 2019, ADV NEUR IN, V32
  • [37] Zareian A., 2020, ARXIV PREPRINT ARXIV
  • [38] Zellers R, 2018, 2018 CONFERENCE ON EMPIRICAL METHODS IN NATURAL LANGUAGE PROCESSING (EMNLP 2018), P93