Skip to content

Analyze high-quality AI papers from SOTA!

Notifications You must be signed in to change notification settings

JayTing511/papers

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

59 Commits
 
 

Repository files navigation

Logo Build Status


论文说明

以下论文均来自人工智能领域SOTA排行榜的精选论文,它们代表学术界最新的研究方向与成果,是AI未来的前进的方向。


精选论文

1. Understanding Back-Translation at Scale

当前在WMT2014英语到德语翻译集上表现最好的机器翻译模型是Transformer Big,其模型在提升中使用的技巧BT(Back-Translation)正是源自于该论文。

点击查看


2. Language Models are Few-Shot Learners

当前在PennTreebank数据集上表现最佳的语言模型是GPT-3,它能够很好的应对包括翻译,问题解答和完形填空等下游任务,其中主要实验过程和结论正源于该论文。

点击查看


3. Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer

从论文的名字可以看出,这正是支持大名鼎鼎的Google T5-3B的源论文,它在SQuAD1.1 dev问答数据集上取得了当前的SOTA水平。

点击查看


4. Fine-grained Sentiment Classification using BERT

当前在SST-5 Fine-grained classification数据集上表现最佳的情感分析模型是BERT-large,其中主要实验过程和结论正源于该论文。

点击查看


5. Object-Contextual Representations for Semantic Segmentation

当前在Cityscapes test数据集上表现最佳的语义分割模型是HRNetV2,其中主要实验过程和结论正源于该论文。

点击查看


6. Fixing the train-test resolution discrepancy: FixEfficientNet

当前在ImageNet数据集上表现最佳的图像分类模型是FixEfficientNet-L2,其中主要实验过程和结论正源于该论文。

点击查看


7. DetectoRS: Detecting Objects with Recursive Feature Pyramid and Switchable Atrous Convolution

当前在COCO test-dev数据集上表现最佳的目标检测模型是DetectoRS (ResNeXt-101-32x4d, multi-scale),其中主要实验过程和结论正源于该论文。

点击查看


8. Generating Long Sequences with Sparse Transformers

当前在ImageNet 64x64数据集上表现最佳的图像生成模型是Sparse Transformer 152M (strided),其中主要实验过程和结论正源于该论文。

点击查看


9. Toward fast and accurate human pose estimation via soft-gated skip connections

当前在MPII Human Pose数据集上表现最佳的姿势估计模型是Soft-gated Skip Connections,其中主要实验过程和结论正源于该论文。

点击查看


10. English Conversational Telephone Speech Recognition by Humans and Machines

当前在Switchboard + Hub500数据集上表现最佳的语音识别模型是ResNet + BiLSTMs acoustic model,其中主要实验过程和结论正源于该论文。

点击查看


其他资源


致谢

感谢以下机构或站点提供的人工智能工具和参考学习资料。所有人可以通过点击下方图片阅读免费电子书籍或访问这些机构的官方站点。

avatar avatar avatar
avatar avatar avatar

About

Analyze high-quality AI papers from SOTA!

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published