In this work, we present a compact, modular framework for constructing novel
recurrent neural architectures. Our basic module is a new generic unit, the
Transition Based Recurrent Unit (TBRU). In addition to hidden layer
activations, TBRUs have discrete state dynamics that allow network connections
to be built dynamically as a function of intermediate activations. By
connecting multiple TBRUs, we can extend and combine commonly used
architectures such as sequence-to-sequence, attention mechanisms, and
re-cursive tree-structured models. A TBRU can also serve as both an encoder for
downstream tasks and as a decoder for its own task simultaneously, resulting in
more accurate multi-task learning. We call our approach Dynamic Recurrent
Acyclic Graphical Neural Networks, or DRAGNN. We show that DRAGNN is
significantly more accurate and efficient than seq2seq with attention for
syntactic dependency parsing and yields more accurate multi-task learning for
extractive summarization tasks.
Beschreibung
[1703.04474] DRAGNN: A Transition-based Framework for Dynamically Connected Neural Networks
%0 Generic
%1 kong2017dragnn
%A Kong, Lingpeng
%A Alberti, Chris
%A Andor, Daniel
%A Bogatyy, Ivan
%A Weiss, David
%D 2017
%K mlnlp neuralnet parsing syntaxnet
%T DRAGNN: A Transition-based Framework for Dynamically Connected Neural
Networks
%U http://arxiv.org/abs/1703.04474
%X In this work, we present a compact, modular framework for constructing novel
recurrent neural architectures. Our basic module is a new generic unit, the
Transition Based Recurrent Unit (TBRU). In addition to hidden layer
activations, TBRUs have discrete state dynamics that allow network connections
to be built dynamically as a function of intermediate activations. By
connecting multiple TBRUs, we can extend and combine commonly used
architectures such as sequence-to-sequence, attention mechanisms, and
re-cursive tree-structured models. A TBRU can also serve as both an encoder for
downstream tasks and as a decoder for its own task simultaneously, resulting in
more accurate multi-task learning. We call our approach Dynamic Recurrent
Acyclic Graphical Neural Networks, or DRAGNN. We show that DRAGNN is
significantly more accurate and efficient than seq2seq with attention for
syntactic dependency parsing and yields more accurate multi-task learning for
extractive summarization tasks.
@misc{kong2017dragnn,
abstract = {In this work, we present a compact, modular framework for constructing novel
recurrent neural architectures. Our basic module is a new generic unit, the
Transition Based Recurrent Unit (TBRU). In addition to hidden layer
activations, TBRUs have discrete state dynamics that allow network connections
to be built dynamically as a function of intermediate activations. By
connecting multiple TBRUs, we can extend and combine commonly used
architectures such as sequence-to-sequence, attention mechanisms, and
re-cursive tree-structured models. A TBRU can also serve as both an encoder for
downstream tasks and as a decoder for its own task simultaneously, resulting in
more accurate multi-task learning. We call our approach Dynamic Recurrent
Acyclic Graphical Neural Networks, or DRAGNN. We show that DRAGNN is
significantly more accurate and efficient than seq2seq with attention for
syntactic dependency parsing and yields more accurate multi-task learning for
extractive summarization tasks.},
added-at = {2018-07-04T20:50:44.000+0200},
author = {Kong, Lingpeng and Alberti, Chris and Andor, Daniel and Bogatyy, Ivan and Weiss, David},
biburl = {https://www.bibsonomy.org/bibtex/20ecd43eab6c4f6f396d5270488055f46/albinzehe},
description = {[1703.04474] DRAGNN: A Transition-based Framework for Dynamically Connected Neural Networks},
interhash = {e4e7f1a6231fe8c120108504fc26709a},
intrahash = {0ecd43eab6c4f6f396d5270488055f46},
keywords = {mlnlp neuralnet parsing syntaxnet},
note = {cite arxiv:1703.04474Comment: 10 pages; Submitted for review to ACL2017},
timestamp = {2018-07-04T20:50:44.000+0200},
title = {DRAGNN: A Transition-based Framework for Dynamically Connected Neural
Networks},
url = {http://arxiv.org/abs/1703.04474},
year = 2017
}