# Self-Attentional Models for Lattice Inputs

arXiv: https://arxiv.org/abs/1906.01617 · HTML: https://mlsanity.com/paper/1906.01617/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 4 Jun 2019.

## Built on by (most cited first, 12 of 12)

- [STEMM: Self-learning with Speech-text Manifold Mixup for Speech Translation](https://mlsanity.com/paper/2203.10426/) (2022)
- [Cascade versus Direct Speech Translation: Do the Differences Still Make a Difference?](https://mlsanity.com/paper/2106.01045/) (2021)
- [Learning Shared Semantic Space for Speech-to-Text Translation](https://mlsanity.com/paper/2105.03095/) (2021)
- [LLaST: Improved End-to-end Speech Translation System Leveraged by Large Language Models](https://mlsanity.com/paper/2407.15415/) (2024)
- [Lattention: Lattice-attention in ASR rescoring](https://mlsanity.com/paper/2111.10157/) (2021)
- [An Empirical Study of Consistency Regularization for End-to-End Speech-to-Text Translation](https://mlsanity.com/paper/2308.14482/) (2023)
- [EEL: Efficiently Encoding Lattices for Reranking](https://mlsanity.com/paper/2306.00947/) (2023)
- [When End-to-End is Overkill: Rethinking Cascaded Speech-to-Text Translation](https://mlsanity.com/paper/2502.00377/) (2025)
- [Improving speech translation by fusing speech and text](https://mlsanity.com/paper/2305.14042/) (2023)
- [Representation Purification for End-to-End Speech Translation](https://mlsanity.com/paper/2412.04266/) (2024)
- [LT-LM: a novel non-autoregressive language model for single-shot lattice rescoring](https://mlsanity.com/paper/2104.02526/) (2021)
- [Confusion2vec 2.0: Enriching Ambiguous Spoken Language Representations with Subwords](https://mlsanity.com/paper/2102.02270/) (2021)

## Abstract (arXiv)

Lattices are an efficient and effective method to encode ambiguity of upstream systems in natural language processing tasks, for example to compactly capture multiple speech recognition hypotheses, or to represent multiple linguistic analyses. Previous work has extended recurrent neural networks to model lattice inputs and achieved improvements in various tasks, but these models suffer from very slow computation speeds. This paper extends the recently proposed paradigm of self-attention to handle lattice inputs. Self-attention is a sequence modeling technique that relates inputs to one another by computing pairwise similarities and has gained popularity for both its strong results and its computational efficiency. To extend such models to handle lattices, we introduce probabilistic reachability masks that incorporate lattice structure into the model and support lattice scores if available. We also propose a method for adapting positional embeddings to lattice structures. We apply the proposed model to a speech translation task and find that it outperforms all examined baselines while being much faster to compute than previous neural lattice models during both training and inference.