Writing Polishment with Simile: Task, Dataset and A Neural Approach

A simile is a figure of speech that directly makes comparison, showing similarities between two different things, e.g. ``Reading papers can be dull sometimes,like watching grass grow". Human writers often interpolate appropriate similes into proper locations the plain text to vivify their writings. However, none existing work has explored neural interpolation, including both locating and generation. In this paper, we propose new task Writing Polishment with Simile (WPS) investigate whether machines are able polish texts as human do. Accordingly, design two-staged Locate&Gen model based on transformer architecture. Our firstly locates where interpolation should happen, then generates location-specific simile. We also release large-scale Chinese (CS) dataset containing 5 million context. The experimental results demonstrate feasibility WPS shed light future research directions towards better automatic polishment.