Skip to Main content Skip to Navigation
Other publications

Hybrid Neural Models For Sequence Modelling: The Best Of Three Worlds

Abstract : We propose a neural architecture with the main characteristics of the most successful neural models of the last years : bidirectional RNNs, encoder-decoder, and the Transformer model. Evaluation on three sequence labelling tasks yields results that are close to the state-of-the-art for all tasks and better than it for some of them, showing the pertinence of this hybrid architecture for this kind of tasks.
Complete list of metadata

Cited literature [51 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-02416404
Contributor : Marco Dinarelli <>
Submitted on : Tuesday, December 17, 2019 - 4:09:20 PM
Last modification on : Thursday, July 1, 2021 - 5:46:02 PM
Long-term archiving on: : Wednesday, March 18, 2020 - 9:17:35 PM

File

2019_arXiv_Seq2Biseq-Transform...
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02416404, version 1

Collections

Citation

Marco Dinarelli, Loïc Grobol. Hybrid Neural Models For Sequence Modelling: The Best Of Three Worlds. 2019. ⟨hal-02416404⟩

Share

Metrics

Record views

46

Files downloads

138