Insertion-based Decoding with Automatically Inferred Generation Order
Conventional neural autoregressive decoding commonly assumes a fixed left-to-right generation order, which may be sub-optimal. In this work, we propose a novel decoding algorithm— InDIGO—which supports flexible sequence generation in arbitrary orders through insertion operation...
Main Authors: | Gu, Jiatao, Liu, Qi, Cho, Kyunghyun |
---|---|
Format: | Article |
Language: | English |
Published: |
The MIT Press
2019-11-01
|
Series: | Transactions of the Association for Computational Linguistics |
Online Access: | https://www.mitpressjournals.org/doi/abs/10.1162/tacl_a_00292 |
Similar Items
-
Automatic inference for higher-order probabilistic programs
by: Paige, Timothy Brooks
Published: (2016) -
Automatic Comma Insertion for Japanese Text Generation
by: Matsubara, Shigeki, et al.
Published: (2010) -
Automatic inference of code transforms for patch generation
by: Long, Fan, et al.
Published: (2019) -
Code generation and simulation of an automatic, flexible QC-LDPC hardware decoder
by: Von Leipzig, Mirko
Published: (2015) -
Voice Encode/Decode System for PCM Insertion and Extraction
by: Laird, Daniel T.
Published: (1994)