Spectral Text Similarity Measures.

Estimating semantic similarity between texts is of vital importance in many areas of natural language processing like information retrieval, question answering, text reuse, or plagiarism detection. Prevalent semantic similarity estimates based on word embeddings are noise sensitive. Thus, small individual term similarities can have in aggregate a considerable influence on the total estimation value. In contrast, the methods proposed here exploit the spectrum of the product of embedding matrices, which leads to increased robustness when compared with conventional methods. We apply these estimate on two tasks, which are the assignment of people to the best matching marketing target group and finding the correct match between sentences belonging to two independent translations of the same novel. The evaluation revealed that our proposed method based on the spectral norm could increase the accuracy compared to several baseline methods in both scenarios.