PaperRobot: Incremental Draft Generation of Scientific Ideas
Q. Wang, L. Huang, Z. Jiang, K. Knight, H. Ji, M. Bansal, und Y. Luan. (2019)cite arxiv:1905.07870Comment: 12 pages. Accepted by ACL 2019 Code and resource will be available at https://github.com/EagleW/PaperRobot.
Zusammenfassung
We present a PaperRobot who performs as an automatic research assistant by
(1) conducting deep understanding of a large collection of human-written papers
in a target domain and constructing comprehensive background knowledge graphs
(KGs); (2) creating new ideas by predicting links from the background KGs, by
combining graph attention and contextual text attention; (3) incrementally
writing some key elements of a new paper based on memory-attention networks:
from the input title along with predicted related entities to generate a paper
abstract, from the abstract to generate conclusion and future work, and finally
from future work to generate a title for a follow-on paper. Turing Tests, where
a biomedical domain expert is asked to compare a system output and a
human-authored string, show PaperRobot generated abstracts, conclusion and
future work sections, and new titles are chosen over human-written ones up to
30%, 24% and 12% of the time, respectively.
Beschreibung
[1905.07870] PaperRobot: Incremental Draft Generation of Scientific Ideas
%0 Generic
%1 wang2019paperrobot
%A Wang, Qingyun
%A Huang, Lifu
%A Jiang, Zhiying
%A Knight, Kevin
%A Ji, Heng
%A Bansal, Mohit
%A Luan, Yi
%D 2019
%K nlp
%T PaperRobot: Incremental Draft Generation of Scientific Ideas
%U http://arxiv.org/abs/1905.07870
%X We present a PaperRobot who performs as an automatic research assistant by
(1) conducting deep understanding of a large collection of human-written papers
in a target domain and constructing comprehensive background knowledge graphs
(KGs); (2) creating new ideas by predicting links from the background KGs, by
combining graph attention and contextual text attention; (3) incrementally
writing some key elements of a new paper based on memory-attention networks:
from the input title along with predicted related entities to generate a paper
abstract, from the abstract to generate conclusion and future work, and finally
from future work to generate a title for a follow-on paper. Turing Tests, where
a biomedical domain expert is asked to compare a system output and a
human-authored string, show PaperRobot generated abstracts, conclusion and
future work sections, and new titles are chosen over human-written ones up to
30%, 24% and 12% of the time, respectively.
@misc{wang2019paperrobot,
abstract = {We present a PaperRobot who performs as an automatic research assistant by
(1) conducting deep understanding of a large collection of human-written papers
in a target domain and constructing comprehensive background knowledge graphs
(KGs); (2) creating new ideas by predicting links from the background KGs, by
combining graph attention and contextual text attention; (3) incrementally
writing some key elements of a new paper based on memory-attention networks:
from the input title along with predicted related entities to generate a paper
abstract, from the abstract to generate conclusion and future work, and finally
from future work to generate a title for a follow-on paper. Turing Tests, where
a biomedical domain expert is asked to compare a system output and a
human-authored string, show PaperRobot generated abstracts, conclusion and
future work sections, and new titles are chosen over human-written ones up to
30%, 24% and 12% of the time, respectively.},
added-at = {2019-05-22T11:49:36.000+0200},
author = {Wang, Qingyun and Huang, Lifu and Jiang, Zhiying and Knight, Kevin and Ji, Heng and Bansal, Mohit and Luan, Yi},
biburl = {https://www.bibsonomy.org/bibtex/251576fe699c698c1563c048f0c755c0c/straybird321},
description = {[1905.07870] PaperRobot: Incremental Draft Generation of Scientific Ideas},
interhash = {b0a8544bf98a9798fb3d04540f37a698},
intrahash = {51576fe699c698c1563c048f0c755c0c},
keywords = {nlp},
note = {cite arxiv:1905.07870Comment: 12 pages. Accepted by ACL 2019 Code and resource will be available at https://github.com/EagleW/PaperRobot},
timestamp = {2019-05-22T11:49:36.000+0200},
title = {PaperRobot: Incremental Draft Generation of Scientific Ideas},
url = {http://arxiv.org/abs/1905.07870},
year = 2019
}