LOCAL CONTEXT INTERACTION-AWARE GLYPH-VECTORS FOR CHINESE SEQUENCE TAGGING

As hieroglyphics, Chinese characters contain rich semantic and glyphs information, which is beneficial to sequence tagging task. However, it's difficult for shallow CNNs architecture to extract glyphs information from character data and implement the contextual interaction of different glyphs information effectively. In this paper, we address these issues by presenting LCIN: a Local Context Interaction-aware Network for glyph-vectors extraction. The network utilizes depthwise separable convolution and attention machine to extract glyphs information from images of Chinese characters. Moreover, we interconnect adjacent attention blocks so that glyphs information can flow within the local context. Experiments on three subtasks for sequence tagging show that our method outperforms other glyph-based models and achieves new SOTA results in a wide range of datasets.