Iterative Edit-Based Unsupervised Sentence Simplification
This work addresses the problem of simplifying text for broader accessibility without requiring parallel training data, offering a more controllable and interpretable method, though it is incremental relative to supervised approaches.
The paper tackles unsupervised sentence simplification by proposing an iterative edit-based model guided by fluency, simplicity, and meaning preservation, achieving results nearly as effective as state-of-the-art supervised approaches on Newsela and WikiLarge datasets.
We present a novel iterative, edit-based approach to unsupervised sentence simplification. Our model is guided by a scoring function involving fluency, simplicity, and meaning preservation. Then, we iteratively perform word and phrase-level edits on the complex sentence. Compared with previous approaches, our model does not require a parallel training set, but is more controllable and interpretable. Experiments on Newsela and WikiLarge datasets show that our approach is nearly as effective as state-of-the-art supervised approaches.