Penn & BGU BabyBERTa+ for Strict-Small BabyLM Challenge

Yahan Yang, Insup Lee, Elior Sulem, Dan Roth

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

The BabyLM Challenge aims at pre-training a language model on a small-scale dataset of inputs intended for children. In this work, we adapted the architecture and masking policy of BabyBERTa (Huebner et al., 2021) to solve the strict-small track of the BabyLM challenge. Our model, Penn & BGU BabyBERTa+, was pre-trained and evaluated on the three benchmarks of the BabyLM Challenge. Experimental results indicate that our model achieves higher or comparable performance in predicting 17 grammatical phenomena, compared to the RoBERTa baseline.

Original languageAmerican English
Title of host publicationCoNLL 2023 - BabyLM Challenge at the 27th Conference on Computational Natural Language Learning, Proceedings
EditorsAlex Warstadt, Aaron Mueller, Leshem Choshen, Ethan Wilcox, Chengxu Zhuang, Juan Ciro, Rafael Mosquera, Bhargavi Paranjabe, Adina Williams, Tal Linzen, Ryan Cotterell
PublisherAssociation for Computational Linguistics (ACL)
Pages86-88
Number of pages3
ISBN (Electronic)9781952148026
StatePublished - 1 Jan 2023
EventBabyLM Challenge at the 27th Conference on Computational Natural Language Learning, CoNLL 2023 - Singapore, Singapore
Duration: 6 Dec 20237 Dec 2023

Publication series

NameCoNLL 2023 - BabyLM Challenge at the 27th Conference on Computational Natural Language Learning, Proceedings

Conference

ConferenceBabyLM Challenge at the 27th Conference on Computational Natural Language Learning, CoNLL 2023
Country/TerritorySingapore
CitySingapore
Period6/12/237/12/23

All Science Journal Classification (ASJC) codes

  • Linguistics and Language
  • Artificial Intelligence
  • Human-Computer Interaction

Fingerprint

Dive into the research topics of 'Penn & BGU BabyBERTa+ for Strict-Small BabyLM Challenge'. Together they form a unique fingerprint.

Cite this