Contextualized entity representations learned by state-of-the-art transformer-based language models (TLMs) like BERT, GPT, T5, etc., leverage the attention mechanism to learn the data context from training data corpus. However, these models do not use the knowledge context. Knowledge context can be understood as semantics about entities and their relationship with neighboring entities in knowledge graphs. We propose a novel and effective technique to infuse knowledge context from multiple knowledge graphs for conceptual and ambiguous entities into TLMs during fine-tuning. It projects knowledge graph embeddings in the homogeneous vector-space, introduces new token-types for entities, aligns entity position ids, and a selective attention mechanism. We take BERT as a baseline model and implement the "Knowledge-Infused BERT" by infusing knowledge context from ConceptNet and WordNet, which significantly outperforms BERT and other recent knowledge-aware BERT variants like ERNIE, SenseBERT, and BERT_CS over eight different subtasks of GLUE benchmark. The KI-BERT-base model even significantly outperforms BERT-large for domain-specific tasks like SciTail and academic subsets of QQP, QNLI, and MNLI.

翻译：由BERT、GPT、T5等最新变压器基语言模型(TLMs)所学的最新变压器语言模型(TLMs)所学的内幕实体表示,利用关注机制从培训数据库中学习数据背景,然而,这些模型并不使用知识背景。知识背景可以理解为实体的语义,以及它们在知识图中与相邻实体的关系。我们提出了一个新颖而有效的技术,在微调过程中,将概念和模棱两可的实体从多种知识图表中引入知识环境,将概念和模棱两可的实体纳入TLMS(TLER)、将知识图嵌入同质矢量空间,为实体引入新的代号类型,统一实体位置标识,以及选择性关注机制。我们将BERT作为基线模型,并通过从概念网和WordNet中引入知识背景环境来实施“知识应用的BERT(GNIE)、SenseBERT(SERT)和BERT-CSLI(GA)数据库模型,例如SQ-NER-IST-ILILA-ILA-BS-ILABS-GLILILA(S-ILILABS-GLILILABS-ILA)的模型,甚至显著的S-G-G-I-LIS-LIS-LIS-LIS-LIS-LIS-S-S-GF-GF-GT-GT-GLIS-GT-GT-GT-GT-GT-GT-GT-GT-GLIS-GT-GT-GT-GLIS-GT-GT-GT-GT-GT-GT-GT-GT-GT-GT-GT-GT-GT-GT-GT-GT-GT-GT-GLIS-GT-GT-GT-G-G-G-GT-G-G-G-G-G-G-G-G-G-G-GT-GT-GT-GT-G-G-G-GT-GT-GT-GT-G