Coarse-grained Candidate Generation and Fine-grained Re-ranking for Chinese Abbreviation Prediction

Correctly predicting abbreviations given the full forms is important in many natural language processing systems. In this paper we propose a two-stage method to find the corresponding abbreviation given its full form. We first use the contextual information given a large corpus to get abbreviation candidates for each full form and get a coarse-grained ranking through graph random walk. This coarse-grained rank list fixes the search space inside the top-ranked candidates. Then we use a similarity sensitive re-ranking strategy which can utilize the features of the candidates to give a fine-grained re-ranking and select the final result. Our method achieves good results and outperforms the state-ofthe-art systems. One advantage of our method is that it only needs weak supervision and can get competitive results with fewer training data. The candidate generation and coarse-grained ranking is totally unsupervised. The re-ranking phase can use a very small amount of training data to get a reasonably good result.