# Specializing Unsupervised Pretraining Models for Word-Level Semantic Similarity

arXiv: https://arxiv.org/abs/1909.02339 · HTML: https://mlsanity.com/paper/1909.02339/

## Key facts

- Cited by 20 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 5 Sep 2019.

## Built on by (most cited first, 20 of 20)

- [AMMUS : A Survey of Transformer-based Pretrained Models in Natural Language Processing](https://mlsanity.com/paper/2108.05542/) (2021)
- [A Survey of Knowledge Enhanced Pre-trained Language Models](https://mlsanity.com/paper/2211.05994/) (2022)
- [Knowledge Enhanced Pretrained Language Models: A Compreshensive Survey](https://mlsanity.com/paper/2110.08455/) (2021)
- [Learning Domain-Specialised Representations for Cross-Lingual Biomedical Entity Linking](https://mlsanity.com/paper/2105.14398/) (2021)
- [ConvFiT: Conversational Fine-Tuning of Pretrained Language Models](https://mlsanity.com/paper/2109.10126/) (2021)
- [Knowledge-in-Context: Towards Knowledgeable Semi-Parametric Language Models](https://mlsanity.com/paper/2210.16433/) (2022)
- [The Life Cycle of Knowledge in Big Language Models: A Survey](https://mlsanity.com/paper/2303.07616/) (2023)
- [Low Anisotropy Sense Retrofitting (LASeR) : Towards Isotropic and Sense Enriched Representations](https://mlsanity.com/paper/2104.10833/) (2021)
- [MESED: A Multi-modal Entity Set Expansion Dataset with Fine-grained Semantic Classes and Hard Negative Entities](https://mlsanity.com/paper/2307.14878/) (2023)
- [K-PLUG: Knowledge-injected Pre-trained Language Model for Natural Language Understanding and Generation in E-Commerce](https://mlsanity.com/paper/2104.06960/) (2021)
- [Embracing Ambiguity: Improving Similarity-oriented Tasks with Contextual Synonym Knowledge](https://mlsanity.com/paper/2211.10997/) (2022)
- [CD2CR: Co-reference Resolution Across Documents and Domains](https://mlsanity.com/paper/2101.12637/) (2021)
- [Mixed Attention Transformer for Leveraging Word-Level Knowledge to Neural Cross-Lingual Information Retrieval](https://mlsanity.com/paper/2109.02789/) (2021)
- [A Survey of Knowledge Enhanced Pre-trained Models](https://mlsanity.com/paper/2110.00269/) (2021)
- [A Survey on Knowledge-Enhanced Pre-trained Language Models](https://mlsanity.com/paper/2212.13428/) (2022)
- [Small Models, Big Impact: Efficient Corpus and Graph-Based Adaptation of Small Multilingual Language Models for Low-Resource Languages](https://mlsanity.com/paper/2502.10140/) (2025)
- [GANTEE: Generative Adversatial Network for Taxonomy Entering Evaluation](https://mlsanity.com/paper/2303.14480/) (2023)
- [Shuo Wen Jie Zi: Rethinking Dictionaries and Glyphs for Chinese Language Pre-training](https://mlsanity.com/paper/2305.18760/) (2023)
- [Massively Multilingual Lexical Specialization of Multilingual Transformers](https://mlsanity.com/paper/2208.01018/) (2022)
- [Sem4SAP: Synonymous Expression Mining From Open Knowledge Graph For Language Model Synonym-Aware Pretraining](https://mlsanity.com/paper/2303.14425/) (2023)

## Abstract (arXiv)

Unsupervised pretraining models have been shown to facilitate a wide range of downstream NLP applications. These models, however, retain some of the limitations of traditional static word embeddings. In particular, they encode only the distributional knowledge available in raw text corpora, incorporated through language modeling objectives. In this work, we complement such distributional knowledge with external lexical knowledge, that is, we integrate the discrete knowledge on word-level semantic similarity into pretraining. To this end, we generalize the standard BERT model to a multi-task learning setting where we couple BERT's masked language modeling and next sentence prediction objectives with an auxiliary task of binary word relation classification. Our experiments suggest that our "Lexically Informed" BERT (LIBERT), specialized for the word-level semantic similarity, yields better performance than the lexically blind "vanilla" BERT on several language understanding tasks. Concretely, LIBERT outperforms BERT in 9 out of 10 tasks of the GLUE benchmark and is on a par with BERT in the remaining one. Moreover, we show consistent gains on 3 benchmarks for lexical simplification, a task where knowledge about word-level semantic similarity is paramount.