Pre-training Language Model Incorporating Domain-specific Heterogeneous Knowledge into A Unified Representation
arXiv:2109.01048
Abstract
Existing technologies expand BERT from different perspectives, e.g. designing different pre-training tasks, different semantic granularities, and different model architectures. Few models consider expanding BERT from different text formats. In this paper, we propose a heterogeneous knowledge language model (\textbf{HKLM}), a unified pre-trained language model (PLM) for all forms of text, including unstructured text, semi-structured text, and well-structured text. To capture the corresponding relations among these multi-format knowledge, our approach uses masked language model objective to learn word knowledge, uses triple classification objective and title matching objective to learn entity knowledge and topic knowledge respectively. To obtain the aforementioned multi-format text, we construct a corpus in the tourism domain and conduct experiments on 5 tourism NLP datasets. The results show that our approach outperforms the pre-training of plain text using only 1/4 of the data. We further pre-train the domain-agnostic HKLM and achieve performance gains on the XNLI dataset.
References in corpus (8)
- Decoupled Weight Decay Regularization
- BioBERT: a pre-trained biomedical language representation model for biomedical text mining
- Domain-Specific Language Model Pretraining for Biomedical Natural Language Processing
- ERNIE: Enhanced Representation through Knowledge Integration
- ClinicalBERT: Modeling Clinical Notes and Predicting Hospital Readmission
- FinBERT: Financial Sentiment Analysis with Pre-trained Language Models
- PatentBERT: Patent Classification with Fine-Tuning a pre-trained BERT Model
- NeurIPS 2020 EfficientQA Competition: Systems, Analyses and Lessons Learned