Overcoming Poor Word Embeddings with Word Definitions

被引:0
|
作者
Malon, Christopher [1 ]
机构
[1] NEC Labs Amer, Princeton, NJ 08540 USA
关键词
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Modern natural language understanding models depend on pretrained subword embeddings, but applications may need to reason about words that were never or rarely seen during pretraining. We show that examples that depend critically on a rarer word are more challenging for natural language inference models. Then we explore how a model could learn to use definitions, provided in natural text, to overcome this handicap. Our model's understanding of a definition is usually weaker than a well-modeled word embedding, but it recovers most of the performance gap from using a completely untrained word.
引用
收藏
页码:288 / 293
页数:6
相关论文
共 50 条
  • [11] isiZulu Word Embeddings
    Dlamini, Sibonelo
    Jembere, Edgar
    Pillay, Anban
    van Niekerk, Brett
    2021 CONFERENCE ON INFORMATION COMMUNICATIONS TECHNOLOGY AND SOCIETY (ICTAS), 2021, : 121 - 126
  • [12] Topical Word Embeddings
    Liu, Yang
    Liu, Zhiyuan
    Chua, Tat-Seng
    Sun, Maosong
    PROCEEDINGS OF THE TWENTY-NINTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, 2015, : 2418 - 2424
  • [13] Dynamic Word Embeddings
    Bamler, Robert
    Mandt, Stephan
    INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 70, 2017, 70
  • [14] Bias in Word Embeddings
    Papakyriakopoulos, Orestis
    Hegelich, Simon
    Serrano, Juan Carlos Medina
    Marco, Fabienne
    FAT* '20: PROCEEDINGS OF THE 2020 CONFERENCE ON FAIRNESS, ACCOUNTABILITY, AND TRANSPARENCY, 2020, : 446 - 457
  • [15] Compressing Word Embeddings
    Andrews, Martin
    NEURAL INFORMATION PROCESSING, ICONIP 2016, PT IV, 2016, 9950 : 413 - 422
  • [16] Relational Word Embeddings
    Camacho-Collados, Jose
    Espinosa-Anke, Luis
    Schockaert, Steven
    57TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2019), 2019, : 3286 - 3296
  • [17] Unsupervised Word Sense Disambiguation Using Word Embeddings
    Moradi, Behzad
    Ansari, Ebrahim
    Zabokrtsky, Zdenek
    PROCEEDINGS OF THE 2019 25TH CONFERENCE OF OPEN INNOVATIONS ASSOCIATION (FRUCT), 2019, : 228 - 233
  • [18] Integrating Word Embeddings into IBM Word Alignment Models
    Anh-Cuong Le
    Tuan-Phong Nguyen
    Quoc-Long Tran
    PROCEEDINGS OF 2018 10TH INTERNATIONAL CONFERENCE ON KNOWLEDGE AND SYSTEMS ENGINEERING (KSE), 2018, : 79 - 84
  • [19] THE DEFINITIVE WORD ON DEFINITIONS
    YOUNGER, I
    ABA JOURNAL, 1986, 72 : 98 - 98
  • [20] Semantic Word Cloud Generation Based on Word Embeddings
    Xu, Jin
    Tao, Yubo
    Lin, Hai
    2016 IEEE PACIFIC VISUALIZATION SYMPOSIUM (PACIFICVIS), 2016, : 239 - 243