Levenshtein Transformer
Jiatao Gu, Changhan Wang, Jake Zhao

TL;DR
The Levenshtein Transformer introduces a flexible sequence generation model using insertion and deletion operations, enabling dynamic length changes and efficient refinement, with applications in translation and summarization.
Contribution
It presents a novel partially autoregressive model that combines insertion and deletion operations for improved flexibility and efficiency in sequence generation and refinement.
Findings
Achieves comparable performance to existing models
Significantly improves efficiency in generation tasks
Demonstrates versatility across different NLP tasks
Abstract
Modern neural sequence generation models are built to either generate tokens step-by-step from scratch or (iteratively) modify a sequence of tokens bounded by a fixed length. In this work, we develop Levenshtein Transformer, a new partially autoregressive model devised for more flexible and amenable sequence generation. Unlike previous approaches, the atomic operations of our model are insertion and deletion. The combination of them facilitates not only generation but also sequence refinement allowing dynamic length changes. We also propose a set of new training techniques dedicated at them, effectively exploiting one as the other's learning signal thanks to their complementary nature. Experiments applying the proposed model achieve comparable performance but much-improved efficiency on both generation (e.g. machine translation, text summarization) and refinement tasks (e.g. automatic…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsNatural Language Processing Techniques · Topic Modeling · Multimodal Machine Learning Applications
MethodsLinear Layer · Absolute Position Encodings · Position-Wise Feed-Forward Layer · Levenshtein Transformer · Residual Connection · Byte Pair Encoding · Dense Connections · Label Smoothing · *Communicated@Fast*How Do I Communicate to Expedia? · Adam
