Increasing The Performance of Cognitively Inspired Data-Efficient Language Models via Implicit Structure Building

Abstract

In this paper, we describe our submission to the BabyLM Challenge 2023 shared task on data-efficient language model (LM) pretraining (Warstadt et al., 2023). We train transformer-based masked language models that incorporate unsupervised predictions about hierarchical sentence structure into the model architecture. Concretely, we use the Structformer architecture (Shen et al., 2021) and variants thereof. StructFormer models have been shown to perform well on unsupervised syntactic induction based on limited pretraining data, and to yield performance improvements over a vanilla transformer architecture (Shen et al., 2021). Evaluation of our models on 39 tasks provided by the BabyLM challenge shows promising improvements of models that integrate a hierarchical bias into the architecture at some particular tasks, even though they fail to consistently outperform the RoBERTa baseline model provided by the shared task organizers on all tasks.

arXiv commentAccepted at the BabyLM shared task at CoNLL 2023

References in corpus (6)

• HuggingFace's Transformers: State-of-the-art Natural Language Processing
• Scaling Laws for Neural Language Models
• SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems
• Findings of the BabyLM Challenge: Sample-Efficient Pretraining on Developmentally Plausible Corpora
• Transformer Grammars: Augmenting Transformer Language Models with Syntactic Inductive Biases at Scale
• Cramming: Training a Language Model on a Single GPU in One Day

Cited by in corpus (1)

• Findings of the BabyLM Challenge: Sample-Efficient Pretraining on Developmentally Plausible Corpora