Multi-Task Learning for Authorship Attribution via Topic Approximation and Competitive Attention

Multi-Task Learning for Authorship Attribution via Topic Approximation and Competitive Attention
复制标题

通过主题近似和竞争性注意力进行作者归属的多任务学习

DOI:
10.1109/access.2019.2957152
复制
发表时间:
2019
期刊:
影响因子:
3.9
通讯作者:
Lizhen Liu
Lizhen Liu
中科院分区:
计算机科学3区
文献类型:
--
作者:
Wei Song;Chen Zhao;Lizhen Liu

文献摘要

参考文献

相似文献

内容与风格的分离是作者归属中的一个基本问题,它反映了作者独立于主题的个人风格。以前的工作往往忽略了这个问题,强加强有力的,但不切实际的假设或人为地确定一组预定义的文体结构。提出了一种基于神经元多任务学习的主题和风格分离方法。我们的目标是分别学习主题和风格的单独表示。除了作者归属作为主要任务之外,我们还引入了一种新的辅助任务主题近似来指导主题表征的学习,主题模型由外部语料库训练而成。此外,我们提出了一个竞争注意力机制和分离-重构约束,以分配不同的和竞争的注意力,以两个任务,以分离的主题和风格尽可能。评价结果表明,提出的多任务学习的方法是有前途的,特别是在跨主题设置。我们发现,主题近似有助于捕捉主题内容和竞争注意有利于主题风格分离。这是令人鼓舞的,因为我们的模型以概率的方式分离主题和风格,不需要人为干预。
Separating content from style is a fundamental problem in authorship attribution to represent topic independent personal style of authors. Previous work often ignores this problem by imposing strong but unrealistic assumptions or artificially determines a set of predefined stylistic structures. This paper proposes to separate topic and style based on neural multi-task learning. Our target is to learn separate representations for topic and style respectively. In addition to authorship attribution as the main task, we introduce a novel auxiliary task topic approximation to guide the learning of topic representations with the topic distributions inferred by topic models, which are trained from external corpus. Moreover, we propose a competitive attention mechanism and a separation-reconstruction constraint to assign different and competitive attentions to two tasks in order to separate topic and style as much as possible. Evaluation results demonstrate that the proposed multi-task learning based method is promising, especially on cross-topic settings. We found that topic approximation can help capture topical content and the competitive attentions benefit topic-style separation. It is encouraging since our model separates topic and style in a probabilistic way and doesn’t require human intervention.
DOI: 10.3115/1609067.1609104
发表时间: 2009-03
期刊: --
影响因子: --
作者:
Jade Goldstein-Stewart;Ransom K. Winder;R. E. Sabin
通讯作者: Jade Goldstein-Stewart;Ransom K. Winder;R. E. Sabin
DOI: --
发表时间: 2011-09
期刊: --
影响因子: --
作者:
R. Menon;Yejin Choi
通讯作者: R. Menon;Yejin Choi
DOI: --
发表时间: 2018-08
期刊: --
影响因子: --
作者:
Kalaivani Sundararajan;D. Woodard
通讯作者: Kalaivani Sundararajan;D. Woodard
DOI: 10.1002/asi.v60:3
发表时间: 2009-03
影响因子: 3.5
作者:
N. Shibata;Y. Kajikawa;Y. Takeda;K. Matsushima
通讯作者: N. Shibata;Y. Kajikawa;Y. Takeda;K. Matsushima
DOI: --
发表时间: 2017-05
期刊: ArXiv
影响因子: --
作者:
T. Shen;Tao Lei;R. Barzilay;T. Jaakkola
通讯作者: T. Shen;Tao Lei;R. Barzilay;T. Jaakkola