Chinese Nested Named Entity Recognition Based on Suffix Labeling

Named entity recognition is a fundamental task in natural language processing, which aims to detect and classify entitiy mentions in texts. When a span corresponding to an entity is contained within another span, this problem is called nested named entity recognition. Different languages have different grammatical and syntactic structure. We will focus on the task of nested named entity recognition in Chinese. Suffixes of Chinese entity mentions usually contain special information - concepts that make them share similar features with original entities. Based on the above point of view, we propose a simple improvement strategy towards span-based approaches and achieve quite excellent experimental results. To incorporate concept information within Chinese named entities, we apply soft labels to all suffixes of each entity mention and combine label smoothing. This enables the model to capture the relationship between entity mentions and concept mentions, thereby enhancing its performance. <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup><sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup>Our code is available at https://github.com/dpj135/cnner-suffix-labeling.