Meaning is defined by the company it keeps. However, company is two-fold: It's based on the identity of tokens and also on their position (topology). We argue that a position-centric perspective is more general and useful. The classic MLM and CLM objectives in NLP are easily phrased as position predictions over the whole vocabulary. Adapting the relative position encoding paradigm in NLP to create relative labels for self-supervised learning, we seek to show superior pre-training judged by performance on downstream tasks.
Modeling Graph Structure via Relative Position for Text Generation from Knowledge Graphs
Center for Information and Language Processing (CIS), LMU Munich(慕尼黑大学语言与信息处理中心);Research Training Group AIPHES and UKP Lab, Technische Universität Darmstadt(达姆施塔特工业大学)
Source Code Summarization with Structural Relative Position Guided Transformer
Harbin Institute of Technology (Shenzhen)(哈尔滨工业大学(深圳));Noah’s Ark Lab, Huawei Technologies(华为技术有限公司诺亚方舟实验室);The Chinese University of Hong Kong(香港中文大学)
NAG: A Unified Native Architecture for Encoder-free Text-Graph Modeling in Language Models
New Laboratory of Pattern Recognition (NLPR), Institute of Automation, Chinese Academy of Sciences, Beijing, China;School of Artificial Intelligence, University of Chinese Academy of Sciences (UCAS), Beijing, China;School of Advanced Interdisciplinary Sciences, UCAS, Beijing, China