Detecting Short Near-Duplicates with Semantic Relations
Yahao He, Jin Chun Gao · 2018
Detecting short near-duplicates is significant in information retrieval system. In this paper, we propose a novel semantics-encoded method to improve word semantic representation learning and apply it on an improved hash method to detect short near-duplicates. We evaluate our model in SICK dataset and result shows our methodology outperforms all previous baseline model. The result indicates the correctness and good efficiency of our methodology.