Revisiting the Markov Property for machine translation
In this paper, we re-examine the Markov property in the context of neural machine translation. We design a Markov Autoregressive Transformer (MAT) and undertake a comprehensive assessment of its performance across four WMT benchmarks. Our findings indicate that MAT with an order larger than 4 can ge...
Saved in:
Main Authors: | DU, Cunxiao, ZHOU, Hao, TU, Zhaopeng, JIANG, Jing |
---|---|
Format: | text |
Language: | English |
Published: |
Institutional Knowledge at Singapore Management University
2024
|
Subjects: | |
Online Access: | https://ink.library.smu.edu.sg/sis_research/8725 https://ink.library.smu.edu.sg/context/sis_research/article/9728/viewcontent/2024.findings_eacl.40_pvoa.pdf |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Singapore Management University |
Language: | English |
Similar Items
-
Joint Learning for Coreference Resolution with Markov Logic
by: SONG, Yang, et al.
Published: (2012) -
Accurately Extracting Coherent Relevant Passages Using Hidden Markov Models
by: JIANG, Jing, et al.
Published: (2005) -
Order-agnostic cross entropy for non-autoregressive machine translation
by: DU, Cunxiao, et al.
Published: (2021) -
Why accountants should embrace machine learning?
by: LEE, Benjamin Huan Zhou, et al.
Published: (2018) -
ngram-OAXE: Phrase-based order-agnostic cross entropy for non-autoregressive machine translation
by: DU, Cunxiao, et al.
Published: (2022)