Lvcudar Blog
  • Home
  • Archives
  • Tags
  • Categories
Lvcudar Blog
  • Home
  • Archives
  • Tags
  • Categories

Read Paper 《Learning to Select, Track, and Generate for Data-to-Text》

2019-07-27
Paper Note NLP

URL: https://github.com/aistairc/sports-reporter

TL;DR

像人一样追踪结构化数据的亮点,并写出摘要。

Read more...

Read Paper 《GA-DAN: Geometry-Aware Domain Adaptation Network for Scene Text Detection and Recognition》

2019-07-24
Paper Note CV

URL: http://arxiv.org/abs/1907.09466

TL;DR

整合外观变换和几何变换,实现多模态学习,更好的适应不同视角的目标检测,并在场景文字检测中验证。

Read more...

Read Paper 《You Write Like You Eat: Stylistic variation as a predictor of social stratification》

2019-07-18
Paper Note NLP

URL: http://arxiv.org/abs/1907.07265 https://github.com/anbasile/social-variation(尚未存在)

TL;DR

ACL 2019 一篇脑洞大开的创意文章。 受不同阶层人们发音不同研究的启发,在更大的数据集上验证不同阶层的人们的写作文体不一样,并给出了重要特征。

Read more...

Read Paper 《How to make a pizza: Learning a compositional layer-based GAN model》

2019-07-13
Paper Note CV

URL: http://pizzagan.csail.mit.edu/

TL;DR

这是一篇生成图片烹饪方法的生成对抗网络,生成制作皮萨过程中每一步在皮萨饼上面加入或移除佐料的图片。

Read more...

Read Paper 《Sequence-to-Sequence Learning as Beam-Search Optimization》

2019-07-09
Paper Note NLP

URL:https://www.aclweb.org/anthology/D16-1137

lua源代码: https://github.com/harvardnlp/BSO

注释

EMNLP2016 最早提出用束搜索(beam search)优化seq2seq测试时效果,一定程度上克服训练与测试间的模型获知信息偏差和损失于指标间的偏差。

Read more...
Prev Next
Powered by Hugo | Theme - Even © 2019 - 2020 lvcudar