[Adapt] 回复： [Seminar] Style Transfer in Text

[Yizhu Liu]

Hi Adapters,


I will give a talk about our work "Controlling Length in Abstractive Summarization Using a Convolutional Neural Network " which is accepted by EMNLP 2018 as a long paper. 
In this paper, we propose an approach to constrain the summary length by extending a convolutional sequence to sequence model. The results show that this approach generates high-quality summaries with user defined length, and outperforms the baselines consistently in terms of ROUGE score, length variations and semantic similarity.



Best,
Yizhu




------------------ 原始邮件 ------------------
发件人: "章梦雪"<lovealice at sjtu.edu.cn>;
发送时间: 2018年9月19日(星期三) 上午9:50
收件人: "Adapt"<adapt at cs.sjtu.edu.cn>;

主题: [Adapt] [Seminar] Style Transfer in Text



Hi Adapters,

For the seminar in this week, I will talk about style transfer in text. Style transfer is the task of rephrasing the text to contain specific stylistic properties without changing the content in the original text. Due to the lack of large amount of parallel data, algorithms dealing with non-parallel dataset will be presented. Besides, evaluation is also the key challenge for this task. Human evaluation has been widely used to measure the performance of models(same for the style transfer in image). However, human evaluation is subjective and non-efficient. We will discuss about several automatic evaluation metrics.

Hope to see you there : )

Time: 17:00 Sept 19
Venue: SEIEE 3-517A

Best, 
Mengxue
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://cs.sjtu.edu.cn/pipermail/adapt/attachments/20180919/0f44a715/attachment.html>