Simple and Effective Paraphrastic Similarity from Parallel Translations
This work addresses the challenge of efficient and effective paraphrastic similarity for cross-lingual applications, offering a novel approach that bypasses time-consuming data preparation steps.
The authors tackled the problem of learning paraphrastic sentence embeddings by proposing a model that learns directly from parallel translations, eliminating the need for intermediate paraphrase corpora. The resulting model outperforms and is significantly faster than state-of-the-art baselines in cross-lingual tasks, with orders of magnitude speed improvements.
We present a model and methodology for learning paraphrastic sentence embeddings directly from bitext, removing the time-consuming intermediate step of creating paraphrase corpora. Further, we show that the resulting model can be applied to cross-lingual tasks where it both outperforms and is orders of magnitude faster than more complex state-of-the-art baselines.